It sounds like Palan­tir is expe­ri­enc­ing some sig­nif­i­cant employ­ee morale prob­lems. Why? Because it turns out Palan­tir’s Inves­tiga­tive Case Man­age­ment, or ICM, sys­tem that is cur­rent­ly being used by Immi­gra­tion and Cus­toms Enforce­ment (ICE) has been used to build pro­files and track undoc­u­ment­ed immi­grants, includ­ing those immi­grant fam­i­lies where chil­dren have been sep­a­rat­ed by par­ents. Palan­tir’s soft­ware is also used to deter­mine tar­gets for arrest. For exam­ple, ICE agents relied on Palantir’s ICM dur­ing a 2017 oper­a­tion that tar­get­ed fam­i­lies of migrant chil­dren. ICE agents were instruct­ed to use ICM to doc­u­ment any inter­ac­tion they have with unac­com­pa­nied chil­dren try­ing to cross the bor­der and they deter­mined the chil­dren’s par­ents or oth­er fam­i­ly mem­bers facil­i­tat­ed smug­gling them across the bor­der, the fam­i­ly mem­bers could be arrest­ed and pros­e­cut­ed for depor­ta­tion. Ear­li­er this month, the ICE unit that car­ried out the recent high-pro­file raid in Mis­sis­sip­pi — where 680 peo­ple were arrest­ed and detained dur­ing a school day, result­ing in hun­dreds of chil­dren be sent home from school to homes with­out their par­ents — uses Palan­tir’s ICM soft­ware. As the fol­low­ing arti­cle describes, Palan­tir was con­tract­ed in 2014 to build this ICM sys­tem that lets agents access dig­i­tal pro­files of peo­ple sus­pect­ed of vio­lat­ing immi­gra­tion laws and orga­nize records about them in one place. The data in the pro­files includes emails, phone records, text mes­sages and data from auto­mat­ic license plate cam­eras so this is poten­tial­ly very inva­sive data­bas­es of infor­ma­tion on the US immi­grant com­mu­ni­ty.

The fact that the ICM sys­tem is now being used to iden­ti­fy the par­ents and chil­dren who end up get­ting sep­a­rat­ed has under­stand­ably result­ed in a num­ber of Palan­tir employ­ees expe­ri­enc­ing crises of con­science. Although Palan­tir’s lead­er­ship has­n’t expe­ri­ence this cri­sis. Quite the oppo­site. As the fol­low­ing arti­cle describes, Palan­tir has in fact used sim­i­lar sto­ries about employ­ee con­cerns at Google over work the Google was doing for the US mil­i­tary as an oppor­tu­ni­ty to bash Google and declare that Palan­tir would­n’t have such con­cerns about con­tro­ver­sial gov­ern­ment work. And more recent­ly, the com­pa­ny just renewed a $42 mil­lion con­tract with ICE and CEO Alex Karp has defend­ed the role Palan­tir plays with ICE dur­ing com­pa­ny town hall meet­ings. In gen­er­al, it appears that Palan­tir is active­ly try­ing to brand itself in Wash­ing­ton DC as the Sil­i­con Val­ley com­pa­ny that won’t suf­fer from moral qualms about the work its con­tract­ed to do (even if many of the employ­ees are actu­al­ly suf­fer­ing moral qualms):

The Wash­ing­ton Post

The war inside Palan­tir: Data-min­ing firm’s ties to ICE under attack by employ­ees

By Dou­glas MacMil­lan and Eliz­a­beth Dwoskin
August 22, 2019

Alex Karp faced a dilem­ma last year, when employ­ees of the data-min­ing com­pa­ny Palan­tir con­front­ed the chief exec­u­tive with their con­cerns over a part­ner­ship with Immi­gra­tion and Cus­toms Enforce­ment, accord­ing to three peo­ple famil­iar with the inci­dent.

Palan­tir pro­vid­ed dig­i­tal pro­fil­ing tools to the fed­er­al agency as it car­ried out Pres­i­dent Trump’s increas­ing­ly con­tro­ver­sial poli­cies for appre­hend­ing and deport­ing undoc­u­ment­ed immi­grants, trou­bling more than 200 employ­ees who signed a let­ter to Karp, the peo­ple said.

End­ing the con­tracts with ICE would risk a back­lash in Wash­ing­ton, where Palan­tir was quick­ly becom­ing a go-to provider of data-min­ing ser­vices to a wide range of fed­er­al agen­cies. Data min­ing is a process of com­pil­ing mul­ti­tudes of infor­ma­tion from dis­parate sources to show pat­terns and rela­tion­ships. Google’s deci­sion, ear­li­er the same year, to end a con­tract with the Pen­ta­gon over pres­sure from its employ­ees had chilled the Inter­net giant’s rela­tion­ships with some gov­ern­ment lead­ers who accused it of betray­ing Amer­i­can inter­ests.

Karp refused to budge. He renewed an ICE con­tract worth up to $42 mil­lion and defend­ed the pro­gram at a com­pa­ny town hall meet­ing, the peo­ple said. In media inter­views and an online ad cam­paign this year, Karp bashed Google for back­ing out of its gov­ern­ment con­tract and sug­gest­ed Palan­tir wouldn’t do the same.

“Sil­i­con Val­ley is telling the aver­age Amer­i­can ‘I will not sup­port your defense needs,’” Karp told an inter­view­er in Jan­u­ary, a quote the com­pa­ny repeat­ed in a recent ad on Twit­ter. Peter Thiel, Palantir’s bil­lion­aire co-founder, echoed that mes­sage at a con­fer­ence last month, when he called Google’s actions “trea­so­nous.”

“Sil­i­con Val­ley is telling the aver­age Amer­i­can ‘I will not sup­port your defense needs’ while sell­ing prod­ucts to coun­tries that are adver­sar­i­al to Amer­i­ca. That is a los­er posi­tion.”

— Palan­tir (@PalantirTech) July 18, 2019

The con­tro­ver­sy around ICE high­lights a ten­sion at the cen­ter of Palantir’s busi­ness, which relies on the U.S. gov­ern­ment for con­tracts and on Sil­i­con Val­ley for tal­ent. As Trump’s poli­cies divide tech work­ers in the large­ly lib­er­al Bay Area, Palan­tir must bal­ance keep­ing work­ers hap­py and pre­serv­ing the trust of its No. 1 cus­tomer.

Palantir’s predica­ment illus­trates the tightrope walk many busi­ness­es must per­form in an age of ris­ing polit­i­cal activism, par­tic­u­lar­ly in Sil­i­con Val­ley, where tech work­ers have staged walk­outs and cir­cu­lat­ed peti­tions to protest col­lab­o­ra­tions with the Trump admin­is­tra­tion. In their respons­es to work­er upris­ings, the lead­ers of Ama­zon, Google, Microsoft and Sales­force have tried to grap­ple with the eth­i­cal con­cerns posed by their employ­ees — Microsoft, for exam­ple, told employ­ees they don’t have to work on mil­i­tary projects if they don’t want to — while mak­ing it clear they want to keep doing busi­ness with the U.S. gov­ern­ment.

So far, Palan­tir has stood firm in its sup­port of the gov­ern­ment, even as employ­ees and activist groups say there is grow­ing evi­dence that Palan­tir lends sup­port to agents whose work vio­lates the civ­il lib­er­ties of undoc­u­ment­ed immi­grants. A work­place raid result­ing in the arrest of 680 migrant work­ers in Mis­sis­sip­pi on Aug. 7 was car­ried out by the unit of ICE that uses Palan­tir soft­ware to inves­ti­gate poten­tial tar­gets and com­pile evi­dence against them.

In anoth­er employ­ee peti­tion this month, more than 60 Palan­tir work­ers asked man­age­ment to redi­rect the prof­its from ICE con­tracts to a non­prof­it char­i­ty, the peo­ple said. The com­pa­ny renewed a sec­ond ICE con­tract on Aug. 19.

In an inter­view with Bloomberg News this week, Karp said the gov­ern­ment should be respon­si­ble for answer­ing dif­fi­cult ques­tions about how tech­nolo­gies may be used to sur­veil cit­i­zens.

“I do not believe that these ques­tions should be decid­ed in Sil­i­con Val­ley by a num­ber of engi­neers at large plat­form com­pa­nies,” Karp said in the inter­view.

...

Found­ed in the patri­ot­ic fer­vor that fol­lowed the Sept. 11, 2001, ter­ror­ist attacks, with $2 mil­lion in seed mon­ey from a CIA incu­ba­tor, Palan­tir has always pro­mot­ed a mis­sion to defend Amer­i­can inter­ests. Fed­er­al author­i­ties rely on its data plat­form to track down ter­ror­ists, insur­gents, drug smug­glers and insid­er traders, records show.

Palantir’s busi­ness has flour­ished since Trump took office, with rev­enue from U.S. gov­ern­ment con­tracts under his first two-and-a-half years in office already sur­pass­ing its total under Pres­i­dent Barack Obama’s entire sec­ond term. The Army con­tract, award­ed in March and poten­tial­ly worth more than $800 mil­lion, marked the first time a Sil­i­con Val­ley com­pa­ny had been cho­sen to lead a defense pro­gram of record, a type of con­tract with a ded­i­cat­ed line of fund­ing from Con­gress.

Many of Palan­tir 2,500 employ­ees have debat­ed the ICE con­tracts in town hall meet­ings, office hall­ways, Slack chan­nels and email threads, accord­ing to cur­rent and for­mer employ­ees, who spoke on the con­di­tion of anonymi­ty because the com­pa­ny bound them to con­fi­den­tial­i­ty agree­ments. Palan­tir employ­ees, called “Palan­tiri­ans,” have tak­en both sides of the issue: Immi­grant employ­ees have writ­ten heart­felt let­ters shar­ing why they are opposed to the ICE con­tracts, while at least one for­mer ICE offi­cial who now works at Palan­tir has defend­ed them, accord­ing to a cur­rent engi­neer at the com­pa­ny.

Employ­ees who sup­port the ICE part­ner­ship believe Palan­tir has helped the agency do more good than bad, includ­ing sup­port­ing mis­sions to appre­hend dan­ger­ous crim­i­nals, accord­ing to two cur­rent employ­ees. But oth­ers have felt deflat­ed by what they see as management’s lack of recep­tiv­i­ty to their con­cerns, two for­mer employ­ees said. A com­pa­ny with a mis­sion to “work for the com­mon good,” accord­ing to recent job list­ings, increas­ing­ly feels to some work­ers like a tool for Trump’s polit­i­cal agen­da.

“There’s a ver­sion of the sto­ry where they are the good guys,” one for­mer employ­ee said. “Every­one wants to pro­tect ser­vice mem­bers from IEDs. Every­one wants to pre­vent human traf­fick­ing. Not every­one can get behind work­ing for ICE to help deport immi­grants.”

‘Sil­i­con Val­ley kids’ earn respect

For years, Palan­tir was viewed skep­ti­cal­ly by Wash­ing­ton insid­ers, who saw the Palo Alto, Calif.-based com­pa­ny as a rag­tag team of tech pro­gram­mers who wore hood­ies and flip-flops to work. They were “a bunch of Sil­i­con Val­ley kids,” said a for­mer gov­ern­ment offi­cial who award­ed Palan­tir a con­tract in 2009.

That image began to change as ser­vice mem­bers deployed in Afghanistan grew impressed with Palantir’s abil­i­ty to quick­ly assim­i­late troves of data into maps and charts, show­ing the move­ments, for exam­ple, of insur­gents across a land­scape and the like­ly posi­tions of impro­vised explo­sive devices. After tri­al runs with the Navy, Army and Spe­cial Forces, sev­er­al top Pen­ta­gon offi­cials saw Palantir’s plat­form as more pow­er­ful and reli­able than com­pet­ing tools sup­plied by long­time gov­ern­ment con­trac­tors such as Raytheon. The com­pa­ny still strug­gled to win defense busi­ness because of a con­tract pro­cure­ment process that heav­i­ly favored incum­bents.

The $800 mil­lion Army con­tract, in which Palan­tir will build the nerve cen­ter of a vast intel­li­gence gath­er­ing net­work, was pos­si­ble only because Palan­tir suc­cess­ful­ly argued in court that the gov­ern­ment was required by law to con­sid­er pur­chas­ing com­mer­cial prod­ucts, instead of only cus­tom ones made by con­tract­ing firms. It won the court case in 2016, under the Oba­ma admin­is­tra­tion, and won the con­tract this past March, amid a blitz of lob­by­ing and rela­tion­ship-build­ing with the Trump admin­is­tra­tion.

Palantir’s most vis­i­ble tie to the White House is Thiel, the company’s out­spo­ken co-founder, chief backer and exec­u­tive chair­man. An avowed lib­er­tar­i­an who has railed against the tech industry’s pre­dom­i­nant­ly lib­er­al pol­i­tics, Thiel fre­quent­ly embraces con­tro­ver­sy. He gained noto­ri­ety for bankrolling a suc­cess­ful law­suit against the news site Gawk­er, lead­ing to its bank­rupt­cy in 2016. (In an inter­view with the New York Times, he said Gawk­er pub­lished arti­cles that were “very painful and par­a­lyz­ing for peo­ple who were tar­get­ed,” adding: “I thought it was worth fight­ing back.”)

Thiel donat­ed $1.2 mil­lion to Trump’s 2016 cam­paign and stumped for him at the Repub­li­can Nation­al Con­ven­tion, argu­ing he was the leader with the most poten­tial to rebuild the Amer­i­can econ­o­my. He was award­ed a spot on Trump’s tran­si­tion team and helped orga­nize the president’s ini­tial out­reach to tech indus­try lead­ers. At a Trump Tow­er sum­mit for tech CEOs on the eve of Trump’s pres­i­den­cy, Karp was invit­ed to rep­re­sent Palan­tir. Flanked by titans of Ama­zon, Microsoft and Google, his was the small­est com­pa­ny by mar­ket val­ue rep­re­sent­ed at the meet­ing.

The investor, who now lives in Los Ange­les, makes rare appear­ances in Wash­ing­ton, but remains in favor with the pres­i­dent, accord­ing to a per­son close to him. Thiel joined Trump and Ora­cle CEO Safra Catz for a pri­vate din­ner at the White House ear­li­er this year, accord­ing to two peo­ple briefed on the meet­ing. The trio dis­cussed tech com­pa­nies includ­ing Google and Ama­zon, and the $10 bil­lion cloud-com­put­ing con­tract for which Ama­zon is com­pet­ing with Ora­cle, one of the peo­ple said. (Amazon’s founder, Jeff Bezos, owns The Wash­ing­ton Post.)

Sev­er­al Thiel asso­ciates have worked in the admin­is­tra­tion, includ­ing on the tran­si­tion teams at the Pen­ta­gon and the Depart­ment of Com­merce. Both agen­cies sub­se­quent­ly award­ed con­tracts to Palan­tir.

The data-min­ing firm paid lob­by­ists $1.7 mil­lion in 2018 to push for laws that would help open the gov­ern­ment pro­cure­ment process to com­mer­cial tech­nol­o­gy providers.

The busi­ness of war

Google’s with­draw­al from the Defense Depart­ment pro­gram called Project Maven in sum­mer 2018 ignit­ed a debate about how U.S. tech giants should bal­ance the eth­i­cal con­cerns of rank-and-file work­ers and the secu­ri­ty inter­ests of the nation. Thou­sands of Google employ­ees signed a peti­tion argu­ing the com­pa­ny “should not be in the busi­ness of war,” but end­ing the arti­fi­cial intel­li­gence part­ner­ship may have risked Amer­i­can lives, for­mer deputy defense sec­re­tary Bob Work said at the time.

With Google pulling out of a Pen­ta­gon part­ner­ship, Palan­tir saw an oppor­tu­ni­ty to tell gov­ern­ment cus­tomers they wouldn’t do the same, said Kara Fred­er­ick, an asso­ciate fel­low at the Cen­ter for a New Amer­i­can Secu­ri­ty.

“They see that there is a gap in the mar­ket for a com­pa­ny that is will­ing to stand up and say, ‘Yes, we are going to help the U.S. gov­ern­ment achieve its ends,’” said Fred­er­ick, who research­es the use of emerg­ing tech­nolo­gies in defense.

Palantir’s lead­ers joined the crit­i­cism of Google. In a CNBC inter­view in Jan­u­ary, Karp said Sil­i­con Val­ley com­pa­nies that refuse to work with the U.S. gov­ern­ment are “bor­der­line craven” and added that he’s hap­py Palan­tir is “not on that side of the debate.” In a speech to the Nation­al Con­ser­vatism Con­fer­ence in July, Thiel claimed, with­out evi­dence, that Google has been “infil­trat­ed by Chi­nese intel­li­gence.”

Thiel didn’t men­tion his own ties to a com­pa­ny that ben­e­fit­ed from Google’s deci­sion to pull out of the Pen­ta­gon deal. Anduril, a defense start-up backed by Thiel’s invest­ment firm, Founders Fund, was recent­ly award­ed a con­tract on Project Maven.

In a tweet this month, Trump respond­ed to Thiel’s alle­ga­tions against Google, call­ing the investor “a great and bril­liant guy who knows this sub­ject bet­ter than any­one.” In a sep­a­rate fol­low-up, admin­is­tra­tion offi­cials said there was no rea­son to sus­pect espi­onage at Google.

In a state­ment, a Google spokes­woman said the com­pa­ny con­tin­ues to work with the Defense Depart­ment in areas such as cyber­se­cu­ri­ty and health care, and does not work with the Chi­nese mil­i­tary.

Raids and depor­ta­tions

Now, Palan­tir is in the crosshairs of activists.

Pro­test­ers from civ­il rights groups, includ­ing Mijente and Jews for Racial and Eco­nom­ic Jus­tice, have gath­ered out­side Palantir’s Man­hat­tan offices in recent weeks to call for an end to the company’s work with ICE. As employ­ees filed into work in the morn­ing, vol­un­teers hand­ed out fliers explain­ing how Palantir’s soft­ware has been used by ICE agents tar­get­ing migrant work­ers.

“Tell man­age­ment that you do not want Palan­tir involved in con­tracts that harm immi­grants,” one fli­er read.

Palan­tir began work­ing with the Depart­ment of Home­land Secu­ri­ty, the agency that over­sees ICE, in 2011. The com­pa­ny was involved in an effort called “Oper­a­tion Fall­en Hero,” which hunt­ed down mem­bers of the Los Zetas drug traf­fick­ing ring believed to have mur­dered an ICE spe­cial agent. Palantir’s soft­ware was used to assim­i­late data from the Drug Enforce­ment Admin­is­tra­tion, FBI and DHS — includ­ing sur­veil­lance images, smug­gling routes and elec­tron­ic com­mu­ni­ca­tions — to quick­ly find leads, records show. The oper­a­tion led to 782 arrests for crim­i­nal vio­la­tions and 634 “non­crim­i­nal immi­gra­tion arrests,” accord­ing to an ICE official’s tes­ti­mo­ny.

In 2014, Palan­tir won a con­tract to build a cen­tral dig­i­tal repos­i­to­ry of records, called an Inves­tiga­tive Case Man­age­ment, or ICM, sys­tem. The ICM sys­tem lets agents access dig­i­tal pro­files of peo­ple sus­pect­ed of vio­lat­ing immi­gra­tion laws and orga­nize records about them in one place, accord­ing to DHS doc­u­ments. These records may include inves­tiga­tive evi­dence such as emails, phone records, text mes­sages and data from auto­mat­ic license plate cam­eras, accord­ing to DHS.

ICE agents relied on Palantir’s ICM sys­tem dur­ing a 2017 oper­a­tion that tar­get­ed fam­i­lies of migrant chil­dren, accord­ing to an ICE doc­u­ment pub­lished in May by Mijente and the Inter­cept, an online news ser­vice. As part of the mis­sion, ICE agents were instruct­ed to use ICM to doc­u­ment any inter­ac­tion they have with unac­com­pa­nied chil­dren try­ing to cross the bor­der. If the agency deter­mined their par­ents or oth­er fam­i­ly mem­bers facil­i­tat­ed smug­gling them across the bor­der, the fam­i­ly mem­bers could be arrest­ed and pros­e­cut­ed for depor­ta­tion, the ICE doc­u­ment said.

Mijente has argued that by sup­port­ing this oper­a­tion, Palan­tir was com­plic­it in Trump’s pol­i­cy of sep­a­rat­ing fam­i­lies of undoc­u­ment­ed immi­grants and plac­ing peo­ple in bor­der deten­tion cen­ters with ques­tion­able con­di­tions. Pri­va­cy rights groups includ­ing the Elec­tron­ic Pri­va­cy Infor­ma­tion Cen­ter have raised con­cerns that ICM and FALCON, anoth­er Palan­tir tool used by ICE, may vio­late the pri­va­cy of the peo­ple tracked by these data­bas­es.

Palan­tir has a con­tract with the divi­sion of ICE called Home­land Secu­ri­ty Inves­ti­ga­tions, or HSI. It does not have a con­tract with anoth­er ICE divi­sion called Enforce­ment and Removal Oper­a­tions, or ERO, the unit that his­tor­i­cal­ly has tak­en the lead on raids and depor­ta­tions of undoc­u­ment­ed immi­grants. Karp has fre­quent­ly brought up this dis­tinc­tion when defend­ing the company’s work with ICE, accord­ing to for­mer employ­ees.

How­ev­er, this month, author­i­ties con­firmed that the raids in Mis­sis­sip­pi were car­ried out by HSI, the divi­sion that uses Palan­tir. It’s not clear to what extent Palantir’s prod­ucts have been used to plan or exe­cute work­place raids. Dur­ing prepa­ra­tions for an ICE raid of 7‑Eleven stores across the coun­try last year, an ICE super­vi­sor instruct­ed agents to use Palantir’s FALCON mobile app “to share info with the com­mand cen­ter about the sub­jects encoun­tered in the stores as well as team loca­tions,” accord­ing to emails pub­lished by WNYC last month.

...

———-
“The war inside Palan­tir: Data-min­ing firm’s ties to ICE under attack by employ­ees” by Dou­glas MacMil­lan and Eliz­a­beth Dwoskin; The Wash­ing­ton Post; 08/22/2019

“End­ing the con­tracts with ICE would risk a back­lash in Wash­ing­ton, where Palan­tir was quick­ly becom­ing a go-to provider of data-min­ing ser­vices to a wide range of fed­er­al agen­cies. Data min­ing is a process of com­pil­ing mul­ti­tudes of infor­ma­tion from dis­parate sources to show pat­terns and rela­tion­ships. Google’s deci­sion, ear­li­er the same year, to end a con­tract with the Pen­ta­gon over pres­sure from its employ­ees had chilled the Inter­net giant’s rela­tion­ships with some gov­ern­ment lead­ers who accused it of betray­ing Amer­i­can inter­ests.”

This is the fun­da­men­tal busi­ness prob­lem Palan­tir faces when con­fronting fun­da­men­tal moral prob­lems: its main cus­tomer is the US fed­er­al gov­ern­ment so if it refus­es a con­tract like the ICE case man­age­ment soft­ware con­tract the com­pa­ny risks the rest of those fed­er­al con­tracts. That’s Palan­tir’s busi­ness mod­el. A busi­ness mod­el that includes build­ing the Inves­tiga­tive Case Man­age­ment (ICM) sys­tem that allows ICE to cre­ate detailed dig­i­tal pro­files on indi­vid­u­als. It’s the kind of pow­er­ful tech­nol­o­gy that all sorts of gov­ern­ment agen­cies might be inter­est­ed ing, and maybe even the Palan­tir’s cor­po­rate clients. Build­ing pow­er­ful pro­files of large num­bers of indi­vid­u­als is a gener­i­cal­ly use­ful capa­bil­i­ty to offer clients. But in the end, it’s the US fed­er­al gov­ern­ment that is Palan­tir’s core client and that’s why the com­pa­ny can’t eas­i­ly dis­miss con­tro­ver­sial con­tracts with agen­cies like ICE and when its tools are being used to break up migrant fam­i­lies:

...
In 2014, Palan­tir won a con­tract to build a cen­tral dig­i­tal repos­i­to­ry of records, called an Inves­tiga­tive Case Man­age­ment, or ICM, sys­tem. The ICM sys­tem lets agents access dig­i­tal pro­files of peo­ple sus­pect­ed of vio­lat­ing immi­gra­tion laws and orga­nize records about them in one place, accord­ing to DHS doc­u­ments. These records may include inves­tiga­tive evi­dence such as emails, phone records, text mes­sages and data from auto­mat­ic license plate cam­eras, accord­ing to DHS.

ICE agents relied on Palantir’s ICM sys­tem dur­ing a 2017 oper­a­tion that tar­get­ed fam­i­lies of migrant chil­dren, accord­ing to an ICE doc­u­ment pub­lished in May by Mijente and the Inter­cept, an online news ser­vice. As part of the mis­sion, ICE agents were instruct­ed to use ICM to doc­u­ment any inter­ac­tion they have with unac­com­pa­nied chil­dren try­ing to cross the bor­der. If the agency deter­mined their par­ents or oth­er fam­i­ly mem­bers facil­i­tat­ed smug­gling them across the bor­der, the fam­i­ly mem­bers could be arrest­ed and pros­e­cut­ed for depor­ta­tion, the ICE doc­u­ment said.

Mijente has argued that by sup­port­ing this oper­a­tion, Palan­tir was com­plic­it in Trump’s pol­i­cy of sep­a­rat­ing fam­i­lies of undoc­u­ment­ed immi­grants and plac­ing peo­ple in bor­der deten­tion cen­ters with ques­tion­able con­di­tions. Pri­va­cy rights groups includ­ing the Elec­tron­ic Pri­va­cy Infor­ma­tion Cen­ter have raised con­cerns that ICM and FALCON, anoth­er Palan­tir tool used by ICE, may vio­late the pri­va­cy of the peo­ple tracked by these data­bas­es.

Palan­tir has a con­tract with the divi­sion of ICE called Home­land Secu­ri­ty Inves­ti­ga­tions, or HSI. It does not have a con­tract with anoth­er ICE divi­sion called Enforce­ment and Removal Oper­a­tions, or ERO, the unit that his­tor­i­cal­ly has tak­en the lead on raids and depor­ta­tions of undoc­u­ment­ed immi­grants. Karp has fre­quent­ly brought up this dis­tinc­tion when defend­ing the company’s work with ICE, accord­ing to for­mer employ­ees.

How­ev­er, this month, author­i­ties con­firmed that the raids in Mis­sis­sip­pi were car­ried out by HSI, the divi­sion that uses Palan­tir. It’s not clear to what extent Palantir’s prod­ucts have been used to plan or exe­cute work­place raids. Dur­ing prepa­ra­tions for an ICE raid of 7‑Eleven stores across the coun­try last year, an ICE super­vi­sor instruct­ed agents to use Palantir’s FALCON mobile app “to share info with the com­mand cen­ter about the sub­jects encoun­tered in the stores as well as team loca­tions,” accord­ing to emails pub­lished by WNYC last month.
...

It’s that busi­ness mod­el that’s built around keep the US fed­er­al gov­ern­ment as a core client that makes it no sur­prise to learn that Alex Karp not only dis­missed the con­cerns of those 200 employ­ees, but Palan­tir recent­ly renewed a con­tract with ICE worth $42 mil­lion. In addi­tion, Thiel has pub­licly attacked Google for back­ing out of a fed­er­al gov­ern­ment con­tract and sug­gest­ed that Google was trea­so­nous (as part of alle­ga­tion that the Chi­nese mil­i­tary had infil­trat­ed Google). And Alex Karp recent­ly gave an inter­view where he shared his view that “I do not believe that these ques­tions should be decid­ed in Sil­i­con Val­ley by a num­ber of engi­neers at large plat­form com­pa­nies.” So the mes­sage from Karp appears to be that Palan­tir aren’t actu­al­ly going to engage in any kind of moral deci­sion-mak­ing when it comes to its con­tracts with fed­er­al gov­ern­ment at all. Not con­sid­er­ing the moral­i­ty of its actions is part of this busi­ness mod­el:

...
Palan­tir pro­vid­ed dig­i­tal pro­fil­ing tools to the fed­er­al agency as it car­ried out Pres­i­dent Trump’s increas­ing­ly con­tro­ver­sial poli­cies for appre­hend­ing and deport­ing undoc­u­ment­ed immi­grants, trou­bling more than 200 employ­ees who signed a let­ter to Karp, the peo­ple said.

...

Karp refused to budge. He renewed an ICE con­tract worth up to $42 mil­lion and defend­ed the pro­gram at a com­pa­ny town hall meet­ing, the peo­ple said. In media inter­views and an online ad cam­paign this year, Karp bashed Google for back­ing out of its gov­ern­ment con­tract and sug­gest­ed Palan­tir wouldn’t do the same.

“Sil­i­con Val­ley is telling the aver­age Amer­i­can ‘I will not sup­port your defense needs,’” Karp told an inter­view­er in Jan­u­ary, a quote the com­pa­ny repeat­ed in a recent ad on Twit­ter. Peter Thiel, Palantir’s bil­lion­aire co-founder, echoed that mes­sage at a con­fer­ence last month, when he called Google’s actions “trea­so­nous.”

“Sil­i­con Val­ley is telling the aver­age Amer­i­can ‘I will not sup­port your defense needs’ while sell­ing prod­ucts to coun­tries that are adver­sar­i­al to Amer­i­ca. That is a los­er posi­tion.”

— Palan­tir (@PalantirTech) July 18, 2019

The con­tro­ver­sy around ICE high­lights a ten­sion at the cen­ter of Palantir’s busi­ness, which relies on the U.S. gov­ern­ment for con­tracts and on Sil­i­con Val­ley for tal­ent. As Trump’s poli­cies divide tech work­ers in the large­ly lib­er­al Bay Area, Palan­tir must bal­ance keep­ing work­ers hap­py and pre­serv­ing the trust of its No. 1 cus­tomer.

...

In an inter­view with Bloomberg News this week, Karp said the gov­ern­ment should be respon­si­ble for answer­ing dif­fi­cult ques­tions about how tech­nolo­gies may be used to sur­veil cit­i­zens.

“I do not believe that these ques­tions should be decid­ed in Sil­i­con Val­ley by a num­ber of engi­neers at large plat­form com­pa­nies,” Karp said in the inter­view.
...

And that ‘amoral con­trac­tor for hire’ atti­tude has clear­ly paid off. In March of this year, Palan­tir was award­ed a mas­sive $800 mil­lion con­tract to devel­op a new intel­li­gence gath­er­ing net­work for the US mil­i­tary. Inter­est­ing­ly, in order to win this con­tract, Palan­tir first had to win a court case that found that the fed­er­al gov­ern­ment is required by law to con­sid­er com­mer­cial­ly avail­able prod­ucts instead of only the cus­tom prod­ucts built by con­tract­ing firms. This 2016 court rul­ing essen­tial­ly forced the mil­i­tary into recon­sid­er­ing its deci­sion to go with the estab­lish­ment con­trac­tor, Raytheon, for this big new con­tract and Palan­tir end­ed up win­ning in that con­test. So giv­en that Palan­tir’s com­mer­cial­ly avail­able soft­ware is pre­sum­ably poten­tial­ly applic­a­ble to a lot more gov­ern­ment agen­cies than cur­rent­ly use it, it’s going to be inter­est­ing to see how many new fed­er­al con­tracts with the US gov­ern­ment the com­pa­ny ends up secur­ing in com­ing years:

...
Found­ed in the patri­ot­ic fer­vor that fol­lowed the Sept. 11, 2001, ter­ror­ist attacks, with $2 mil­lion in seed mon­ey from a CIA incu­ba­tor, Palan­tir has always pro­mot­ed a mis­sion to defend Amer­i­can inter­ests. Fed­er­al author­i­ties rely on its data plat­form to track down ter­ror­ists, insur­gents, drug smug­glers and insid­er traders, records show.

Palantir’s busi­ness has flour­ished since Trump took office, with rev­enue from U.S. gov­ern­ment con­tracts under his first two-and-a-half years in office already sur­pass­ing its total under Pres­i­dent Barack Obama’s entire sec­ond term. The Army con­tract, award­ed in March and poten­tial­ly worth more than $800 mil­lion, marked the first time a Sil­i­con Val­ley com­pa­ny had been cho­sen to lead a defense pro­gram of record, a type of con­tract with a ded­i­cat­ed line of fund­ing from Con­gress.

...

The $800 mil­lion Army con­tract, in which Palan­tir will build the nerve cen­ter of a vast intel­li­gence gath­er­ing net­work, was pos­si­ble only because Palan­tir suc­cess­ful­ly argued in court that the gov­ern­ment was required by law to con­sid­er pur­chas­ing com­mer­cial prod­ucts, instead of only cus­tom ones made by con­tract­ing firms. It won the court case in 2016, under the Oba­ma admin­is­tra­tion, and won the con­tract this past March, amid a blitz of lob­by­ing and rela­tion­ship-build­ing with the Trump admin­is­tra­tion.

...

The data-min­ing firm paid lob­by­ists $1.7 mil­lion in 2018 to push for laws that would help open the gov­ern­ment pro­cure­ment process to com­mer­cial tech­nol­o­gy providers.
...

So Palan­tir is going to be even more deeply embed­ded into the US nation­al secu­ri­ty state and mil­i­tary fol­low­ing the com­ple­tion of this new giant Army con­tract to build the nerve cen­ter of a vast intel­li­gence gath­er­ing net­work. What kinds of giant data­bas­es of per­son­al pro­files might this con­tract involve?

And since Palan­tir’s case man­age­ment soft­ware (ICM) that allows for the build­ing of detailed pro­files on large num­bers of peo­ple is one of the main prod­ucts ICE is inter­est­ed in, and pre­sum­ably a lot of oth­er gov­ern­ment agen­cies too, it’s worth recall­ing that the PROMIS mega-scan­dal involved bugged com­mer­cial case man­age­ment soft­ware also devel­oped in coop­er­a­tion with the US gov­ern­ment. It’s espe­cial­ly notable since Palan­tir has oth­er cor­po­rate clients too, as was the case with PROMIS. And, of course, there’s the whole PRISM saga that makes it abun­dant­ly clear Palan­tir is hap­py to assist with spy­ing. In oth­er words, if we were to see a repeat of PROMIS in the mod­ern age, it’s a good bet Palan­tir will be involved. At a min­i­mum, we know the com­pa­ny won’t have any moral qualms about being the next PROMIS.

Here’s the lat­est exam­ple of the GOP’s ongo­ing and grow­ing efforts to ‘work the refs’ in the media and tech indus­try. We’ve already seen how the laugh­able claims of anti-con­ser­v­a­tive bias waged against social media com­pa­nies have become a cen­tral part of the core right-wing strat­e­gy of get­ting favor­able social media treat­ment and ensur­ing the plat­forms remain viable out­lets for right-wing dis­in­for­ma­tion cam­paigns. Now there appears to be a sig­nif­i­cant fund-rais­ing effort to finance a project ded­i­cat­ed to research­ing the past of jour­nal­ists work­ing for vir­tu­al­ly all major main­stream new out­lets, includ­ing their past social media post­ings, and find any­thing that can be embar­rass­ing. The effort is being led by Arthur Schwartz, a Steve Ban­non ally who is described as Don­ald Trump Jr’s “fix­er”.

But it get more devi­ous: this group is claim­ing that they aren’t just going to engage in deep oppo­si­tion research of jour­nal­ists who report things crit­i­cal of Trump. They are also going to be look­ing into the fam­i­ly mem­bers of jour­nal­ists who hap­pen to be active in pol­i­tics and any­one else who works at a media orga­ni­za­tion crit­i­cal of Trump. And any lib­er­al activists of oth­er oppo­nents of Trump will also be sub­ject to this oppo­si­tion research cam­paign. In oth­er words, pret­ty much any­one who does­n’t sup­port Trump and their fam­i­ly mem­bers will be sub­ject to this oppo­si­tion research.

The group has already released dam­ag­ing anti-Semit­ic old tweets from a New York Times edi­tor and a CNN edi­tor. The New York Times edi­tor wrote the tweets while he was in col­lege. The CNN edi­tor wrote them while he was a 15 and 16 year old grow­ing up in Egypt. It under­scores how, after more than a decade of wide­spread social media usage, we now have a large num­ber of peo­ple work­ing in media who were teens clue­less­ly tweet­ing away years ago and now all that old teenage-gen­er­at­ed con­tent is avail­able for use by this net­work.

We’re told by for­mer Ban­non-ally Sam Nun­berg that part of the motive of this oper­a­tion is revenge. Specif­i­cal­ly, revenge against the media for its depic­tion of Trump as a racist. Yep. It’s all part of the gener­ic ‘no, you’re the real racist’ meme that we so often hear these days. But while revenge is the stat­ed goal of this oper­a­tion, it’s also clear­ly part of a media intim­i­da­tion cam­paign as evi­denced by the fact that they are being very out in the open out this:

The New York Times

Trump Allies Tar­get Jour­nal­ists Over Cov­er­age Deemed Hos­tile to White House
The oper­a­tion has com­piled dossiers of poten­tial­ly embar­rass­ing social media posts and oth­er pub­lic state­ments by hun­dreds of peo­ple who work at promi­nent news orga­ni­za­tions.

By Ken­neth P. Vogel and Jere­my W. Peters

Pub­lished Aug. 25, 2019
Updat­ed Aug. 26, 2019

WASHINGTON — A loose net­work of con­ser­v­a­tive oper­a­tives allied with the White House is pur­su­ing what they say will be an aggres­sive oper­a­tion to dis­cred­it news orga­ni­za­tions deemed hos­tile to Pres­i­dent Trump by pub­li­ciz­ing dam­ag­ing infor­ma­tion about jour­nal­ists.

It is the lat­est step in a long-run­ning effort by Mr. Trump and his allies to under­cut the influ­ence of legit­i­mate news report­ing. Four peo­ple famil­iar with the oper­a­tion described how it works, assert­ing that it has com­piled dossiers of poten­tial­ly embar­rass­ing social media posts and oth­er pub­lic state­ments by hun­dreds of peo­ple who work at some of the country’s most promi­nent news orga­ni­za­tions.

The group has already released infor­ma­tion about jour­nal­ists at CNN, The Wash­ing­ton Post and The New York Times — three out­lets that have aggres­sive­ly inves­ti­gat­ed Mr. Trump — in response to report­ing or com­men­tary that the White House’s allies con­sid­er unfair to Mr. Trump and his team or harm­ful to his re-elec­tion prospects.

Oper­a­tives have close­ly exam­ined more than a decade’s worth of pub­lic posts and state­ments by jour­nal­ists, the peo­ple famil­iar with the oper­a­tion said. Only a frac­tion of what the net­work claims to have uncov­ered has been made pub­lic, the peo­ple said, with more to be dis­closed as the 2020 elec­tion heats up. The research is said to extend to mem­bers of jour­nal­ists’ fam­i­lies who are active in pol­i­tics, as well as lib­er­al activists and oth­er polit­i­cal oppo­nents of the pres­i­dent.

It is not pos­si­ble to inde­pen­dent­ly assess the claims about the quan­ti­ty or poten­tial sig­nif­i­cance of the mate­r­i­al the pro-Trump net­work has assem­bled. Some involved in the oper­a­tion have his­to­ries of blus­ter and exag­ger­a­tion. And those will­ing to describe its tech­niques and goals may be try­ing to intim­i­date jour­nal­ists or their employ­ers.

But the mate­r­i­al pub­li­cized so far, while in some cas­es stripped of con­text or pre­sent­ed in mis­lead­ing ways, has proved authen­tic, and much of it has been pro­fes­sion­al­ly harm­ful to its tar­gets.

It is clear from the cas­es to date that among the cen­tral play­ers in the oper­a­tion is Arthur Schwartz, a com­bat­ive 47-year-old con­ser­v­a­tive con­sul­tant who is a friend and infor­mal advis­er to Don­ald Trump Jr., the president’s eldest son. Mr. Schwartz has worked with some of the right’s most aggres­sive oper­a­tives, includ­ing the for­mer Trump advis­er Stephen K. Ban­non.

“If the @nytimes thinks this set­tles the mat­ter we can expose a few of their oth­er big­ots,” Mr. Schwartz tweet­ed on Thurs­day in response to an apolo­getic tweet from a Times jour­nal­ist whose anti-Semit­ic social media posts had just been revealed by the oper­a­tion. “Lots more where this came from.”

The infor­ma­tion unearthed by the oper­a­tion has been com­ment­ed on and spread by offi­cials inside the Trump admin­is­tra­tion and re-elec­tion cam­paign, as well as con­ser­v­a­tive activists and right-wing news out­lets such as Bre­it­bart News. In the case of the Times edi­tor, the news was first pub­lished by Bre­it­bart, imme­di­ate­ly ampli­fied on Twit­ter by Don­ald Trump Jr. and, among oth­ers, Kat­ri­na Pier­son, a senior advis­er to the Trump cam­paign, and quick­ly became the sub­ject of a Bre­it­bart inter­view with Stephanie Grisham, the White House press sec­re­tary and com­mu­ni­ca­tions direc­tor.

The White House press office said that nei­ther the pres­i­dent nor any­one in the White House was involved in or aware of the oper­a­tion, and that nei­ther the White House nor the Repub­li­can Nation­al Com­mit­tee was involved in fund­ing it.

The Trump cam­paign said it was unaware of, and not involved in, the effort, but sug­gest­ed that it served a wor­thy pur­pose. “We know noth­ing about this, but it’s clear that the media has a lot of work to do to clean up its own house,” said Tim Mur­taugh, the campaign’s com­mu­ni­ca­tions direc­tor.

The cam­paign is con­sis­tent with Mr. Trump’s long-run­ning effort to dele­git­imize crit­i­cal report­ing and brand the news media as an “ene­my of the peo­ple.” The pres­i­dent has relent­less­ly sought to dimin­ish the cred­i­bil­i­ty of news orga­ni­za­tions and cast them as polit­i­cal­ly moti­vat­ed oppo­nents.

Jour­nal­ism, he said in a tweet last week, is “noth­ing more than an evil pro­pa­gan­da machine for the Demo­c­rat Par­ty.”

The oper­a­tion has com­piled social media posts from Twit­ter, Face­book and Insta­gram, and stored images of the posts that can be pub­li­cized even if the user deletes them, said the peo­ple famil­iar with the effort. One claimed that the oper­a­tion had unearthed poten­tial­ly “fire­able” infor­ma­tion on “sev­er­al hun­dred” peo­ple.

“I am sure there will be more scalps,” said Sam Nun­berg, a for­mer aide to Mr. Trump who is a friend of Mr. Schwartz.

Mr. Nun­berg and oth­ers who are famil­iar with the cam­paign described it as meant to expose what they see as the hypocrisy of main­stream news out­lets that have report­ed on the president’s inflam­ma­to­ry lan­guage regard­ing race.

“Two can play at this game,” he said. “The media has long tar­get­ed Repub­li­cans with deep dives into their social media, look­ing to car­i­ca­ture all con­ser­v­a­tives and Trump vot­ers as racists.”

But using jour­nal­is­tic tech­niques to tar­get jour­nal­ists and news orga­ni­za­tions as ret­ri­bu­tion for — or as a warn­ing not to pur­sue — cov­er­age crit­i­cal of the pres­i­dent is fun­da­men­tal­ly dif­fer­ent from the well-estab­lished role of the news media in scru­ti­niz­ing peo­ple in posi­tions of pow­er.

“If it’s clear­ly retal­ia­to­ry, it’s clear­ly an attack, it’s clear­ly not jour­nal­ism,” said Leonard Down­ie Jr., who was the exec­u­tive edi­tor of The Post from 1991 to 2008. Ten­sion between a pres­i­dent and the news media that cov­ers him is noth­ing new, Mr. Down­ie added. But an orga­nized, wide-scale polit­i­cal effort to inten­tion­al­ly humil­i­ate jour­nal­ists and oth­ers who work for media out­lets is.

“It’s one thing for Spiro Agnew to call every­one in the press ‘nat­ter­ing nabobs of neg­a­tivism,’” he said, refer­ring to the for­mer vice president’s famous cri­tique of how jour­nal­ists cov­ered Pres­i­dent Richard M. Nixon. “And anoth­er thing to inves­ti­gate indi­vid­u­als in order to embar­rass them pub­licly and jeop­ar­dize their employ­ment.”

...

The oper­a­tion is tar­get­ing the news media by using one of the most effec­tive weapons of polit­i­cal com­bat — deep and labo­ri­ous research into the pub­lic records of oppo­nents to find con­tra­dic­tions, con­tro­ver­sial opin­ions or tox­ic affil­i­a­tions. The lib­er­al group Media Mat­ters for Amer­i­ca helped pio­neer close scruti­ny of pub­lic state­ments by con­ser­v­a­tive media per­son­al­i­ties.

The con­ser­v­a­tive oper­a­tive James O’Keefe has twist­ed that con­cept in ways incon­sis­tent with tra­di­tion­al jour­nal­is­tic ethics, using false iden­ti­ties, elab­o­rate cov­er sto­ries and under­cov­er videos to entrap jour­nal­ists and pub­li­cize embar­rass­ing state­ments, often in mis­lead­ing ways, to under­cut the cred­i­bil­i­ty of what he con­sid­ers news media biased in favor of lib­er­als.

In the case of the pro-Trump net­work, research into jour­nal­ists is being deployed for the polit­i­cal ben­e­fit of the White House. It is tar­get­ing not only high-pro­file jour­nal­ists who chal­lenge the admin­is­tra­tion, but also any­one who works for any news orga­ni­za­tion that mem­bers of the net­work see as hos­tile to Mr. Trump, no mat­ter how tan­gen­tial that job may be to the cov­er­age of his pres­i­den­cy. And it is being used explic­it­ly as ret­ri­bu­tion for cov­er­age.

Some reporters have been warned that they or their news orga­ni­za­tions could be tar­gets, cre­at­ing the impres­sion that the cam­paign is intend­ed in part to deter them from aggres­sive cov­er­age as well as to inflict pun­ish­ment after an arti­cle has been pub­lished.

Trained as a lawyer, Mr. Schwartz has endeared him­self to mem­bers of the president’s fam­i­ly by becom­ing one of their most aggres­sive defend­ers, known for bad­ger­ing and threat­en­ing reporters and oth­ers he believes have wronged the Trumps.

He has pub­licly gone after Repub­li­cans he views as dis­loy­al, includ­ing the for­mer White House chief of staff Reince Priebus, about whom he admit­ted spread­ing an unsub­stan­ti­at­ed rumor. He has called him­self a “troll on Twit­ter,” which is where he has boast­ed of being aware of, or hav­ing access to, dam­ag­ing infor­ma­tion on dozens of jour­nal­ists at CNN and The Times that could be deployed if those out­lets ran afoul of Mr. Trump or his allies.

The operation’s tac­tics were on dis­play last week, seem­ing­ly in response to two pieces in The Times that angered Mr. Trump’s allies. The paper’s edi­to­r­i­al board pub­lished an edi­to­r­i­al on Wednes­day accus­ing Mr. Trump of foment­ing anti-Semi­tism, and the news­room pub­lished a pro­file on Thurs­day morn­ing of Ms. Grisham, the new White House press sec­re­tary, which includ­ed unflat­ter­ing details about her employ­ment his­to­ry.

One per­son involved in the effort said the pro-Trump forces, aware ahead of time about the cov­er­age of Ms. Grisham, were pre­pared to respond. Ear­ly Thurs­day morn­ing, soon after the pro­file appeared online, Bre­it­bart News pub­lished an arti­cle that doc­u­ment­ed anti-Semit­ic and racist tweets writ­ten a decade ago by Tom Wright-Pier­san­ti, who was in col­lege at the time and has since become an edi­tor on the Times’ pol­i­tics desk. The Times said it was review­ing the mat­ter and con­sid­ered the posts “a clear vio­la­tion of our stan­dards.”

Mr. Schwartz tweet­ed a link to the Bre­it­bart piece before 7 a.m., which Don­ald Trump Jr. retweet­ed to his 3.8 mil­lion fol­low­ers — the first of about two dozen times that the president’s son shared the arti­cle or its con­tents. Oth­er promi­nent Repub­li­cans, includ­ing Sen­a­tor Ted Cruz of Texas, joined in high­light­ing the report.

Breitbart’s arti­cle quot­ed sev­er­al peo­ple or groups with close ties to Mr. Schwartz, includ­ing Richard Grenell, Mr. Trump’s ambas­sador to Ger­many, and the Zion­ist Orga­ni­za­tion of Amer­i­ca. It was writ­ten by the site’s Wash­ing­ton polit­i­cal edi­tor, Matthew Boyle, whose rela­tion­ship with Mr. Schwartz start­ed when Mr. Ban­non ran the web­site.

Mr. Boyle’s arti­cle includ­ed a ref­er­ence to the Times pro­file of Ms. Grisham, which it char­ac­ter­ized as “attack­ing White House Press Sec­re­tary Stephanie Grisham.” Mr. Wright-Pier­san­ti was unin­volved in the edit­ing of the arti­cle about Ms. Grisham.

The tweets revealed in the Bre­it­bart arti­cle quick­ly spread to oth­er con­ser­v­a­tive out­lets favored by the pres­i­dent and his allies, includ­ing the radio shows of Rush Lim­baugh and Mark Levin.

Mr. Wright-Pier­san­ti apol­o­gized on Twit­ter on Thurs­day morn­ing and delet­ed offen­sive tweets. Mr. Schwartz then issued his warn­ing that he had fur­ther dam­ag­ing infor­ma­tion about Times employ­ees.

Mr. Wright-Pier­san­ti, 32, said the tweets, post­ed when he was a col­lege stu­dent with a Twit­ter fol­low­ing con­sist­ing most­ly of per­son­al acquain­tances, were “my lame attempts at edgy humor to try to get a rise out of my friends.”

But he said “they’re not fun­ny, they’re clear­ly offen­sive,” adding, “I feel deep shame for them, and I am tru­ly, hon­est­ly sor­ry that I wrote these.”

He said he had for­got­ten about the tweets as he start­ed a career in jour­nal­ism.

“For my gen­er­a­tion, the gen­er­a­tion that came of age in the inter­net, all the youth­ful mis­takes that you made get pre­served in dig­i­tal amber, and no mat­ter how much you change and mature and grow up, it’s always out there, wait­ing to be dis­cov­ered,” Mr. Wright-Pier­san­ti said.

Like Mr. Wright-Pier­san­ti, oth­er tar­gets of the pro-Trump net­work have been young peo­ple who grew up with social media and wrote the posts in ques­tion when they were in their teens or ear­ly 20s, in most cas­es before they became pro­fes­sion­al jour­nal­ists.

A week after a White House reporter for CNN sparred with Mr. Trump dur­ing a news con­fer­ence, Mr. Schwartz high­light­ed a tweet by the reporter from 2011, when the reporter was in col­lege, that used an anti-gay slur. Oth­er sim­i­lar tweets quick­ly sur­faced, and the reporter apol­o­gized, though Mr. Schwartz has con­tin­ued to antag­o­nize the reporter on Twit­ter.

In recent months, Mr. Schwartz high­light­ed a near­ly decade-old tweet in which a reporter for The Post had repeat­ed in an ambigu­ous man­ner a slur used by a politi­cian.

In March, Mr. Schwartz tweet­ed a link to an arti­cle from Bre­it­bart, writ­ten by Mr. Boyle, about a reporter from Busi­ness Insid­er whose Insta­gram account includ­ed anti-Trump ref­er­ences and a pho­to­graph of the reporter demon­strat­ing against the pres­i­dent.

In July, around the time CNN pub­lished an arti­cle expos­ing old posts by a Trump appointee spread­ing sug­ges­tions that Barack Oba­ma was a Mus­lim whose loy­al­ty to the Unit­ed States was in ques­tion, Mr. Schwartz resur­faced anti-Semit­ic tweets from 2011 by a CNN pho­to edi­tor. Mr. Schwartz sug­gest­ed that a CNN reporter who spe­cial­izes in unearthing prob­lem­at­ic archival con­tent should “look into the social media activ­i­ties of your employ­ees.”

The tweets became the basis for sev­er­al arti­cles in con­ser­v­a­tive news out­lets and hun­dreds of tweets from con­ser­v­a­tives tar­get­ing the pho­to edi­tor, Mohammed Elshamy, which did not stop even after he resigned under pres­sure from CNN and apol­o­gized.

“It felt like a coor­di­nat­ed attack,” said Mr. Elshamy, who said he had received death threats. “It was over­whelm­ing.”

Mr. Elshamy, who is now 25, said he post­ed the tweets when he was 15 and 16 years old, grow­ing up in Egypt, when he was still learn­ing Eng­lish and did not ful­ly grasp the mean­ing of the words.

“I was repeat­ing slo­gans heard on the streets dur­ing a high­ly emo­tion­al time in my nation’s his­to­ry,” he said. “I believe that my sub­se­quent work and views over the years redeems for the mis­takes I made as a kid.”

While he said he under­stands “the sever­i­ty and harm of my com­ments,” he ques­tioned the moti­va­tion of the cam­paign that cost him his job. “It is a very dirty tac­tic that they are using to cause as much harm as they can to any­one who is affil­i­at­ed with these media out­lets,” he said. “It actu­al­ly feels like a com­pe­ti­tion and every ter­mi­na­tion or vil­i­fi­ca­tion is a point for them.”

Mr. Ban­non, at the time the head of Bre­it­bart, over­saw the site’s efforts in 2015 to attack Meg­yn Kel­ly, then of Fox News, after she called out Mr. Trump for tweets dis­parag­ing women as “fat pigs,” “dogs” and “slobs.” In an inter­view, he said the work that Mr. Schwartz was under­tak­ing should be seen as a sign that Mr. Trump’s sup­port­ers were com­mit­ted to exe­cut­ing a frontal assault on news media they con­sid­ered adver­sar­i­al.

“A cul­ture war is a war,” he said. “There are casu­al­ties in war. And that’s what you’re see­ing.”

———-
“Trump Allies Tar­get Jour­nal­ists Over Cov­er­age Deemed Hos­tile to White House” by Ken­neth P. Vogel and Jere­my W. Peters; The New York Times; 08/25/2019

“Oper­a­tives have close­ly exam­ined more than a decade’s worth of pub­lic posts and state­ments by jour­nal­ists, the peo­ple famil­iar with the oper­a­tion said. Only a frac­tion of what the net­work claims to have uncov­ered has been made pub­lic, the peo­ple said, with more to be dis­closed as the 2020 elec­tion heats up. The research is said to extend to mem­bers of jour­nal­ists’ fam­i­lies who are active in pol­i­tics, as well as lib­er­al activists and oth­er polit­i­cal oppo­nents of the pres­i­dent.”

Do you sup­port Trump? Nope? Well, get ready for oppo­si­tion research con­duct­ed on you. And this is all being framed as ‘revenge’ against Trump’s oppo­nents for por­tray­ing him, and/or por­tray­ing his sup­port­ers, as racist. This is pre­sum­ably how this kind of intim­i­da­tion cam­paign will be sold to the right-wing audiences...as a ‘we’re fight­ing for you and your hon­or!’ oper­a­tion:

...
Mr. Nun­berg and oth­ers who are famil­iar with the cam­paign described it as meant to expose what they see as the hypocrisy of main­stream news out­lets that have report­ed on the president’s inflam­ma­to­ry lan­guage regard­ing race.

“Two can play at this game,” he said. “The media has long tar­get­ed Repub­li­cans with deep dives into their social media, look­ing to car­i­ca­ture all con­ser­v­a­tives and Trump vot­ers as racists.”

...

In the case of the pro-Trump net­work, research into jour­nal­ists is being deployed for the polit­i­cal ben­e­fit of the White House. It is tar­get­ing not only high-pro­file jour­nal­ists who chal­lenge the admin­is­tra­tion, but also any­one who works for any news orga­ni­za­tion that mem­bers of the net­work see as hos­tile to Mr. Trump, no mat­ter how tan­gen­tial that job may be to the cov­er­age of his pres­i­den­cy. And it is being used explic­it­ly as ret­ri­bu­tion for cov­er­age.

Some reporters have been warned that they or their news orga­ni­za­tions could be tar­gets, cre­at­ing the impres­sion that the cam­paign is intend­ed in part to deter them from aggres­sive cov­er­age as well as to inflict pun­ish­ment after an arti­cle has been pub­lished.
...

And the guy behind, Arthur Schwartz, is both an infor­mal advis­er to Trump Jr. with a his­to­ry of work­ing with Steve Ban­non. As Ban­non describes it, the peo­ple tar­get­ed by this are just casu­al­ties in a cul­ture war:

...
It is clear from the cas­es to date that among the cen­tral play­ers in the oper­a­tion is Arthur Schwartz, a com­bat­ive 47-year-old con­ser­v­a­tive con­sul­tant who is a friend and infor­mal advis­er to Don­ald Trump Jr., the president’s eldest son. Mr. Schwartz has worked with some of the right’s most aggres­sive oper­a­tives, includ­ing the for­mer Trump advis­er Stephen K. Ban­non.

“If the @nytimes thinks this set­tles the mat­ter we can expose a few of their oth­er big­ots,” Mr. Schwartz tweet­ed on Thurs­day in response to an apolo­getic tweet from a Times jour­nal­ist whose anti-Semit­ic social media posts had just been revealed by the oper­a­tion. “Lots more where this came from.”

...

Mr. Ban­non, at the time the head of Bre­it­bart, over­saw the site’s efforts in 2015 to attack Meg­yn Kel­ly, then of Fox News, after she called out Mr. Trump for tweets dis­parag­ing women as “fat pigs,” “dogs” and “slobs.” In an inter­view, he said the work that Mr. Schwartz was under­tak­ing should be seen as a sign that Mr. Trump’s sup­port­ers were com­mit­ted to exe­cut­ing a frontal assault on news media they con­sid­ered adver­sar­i­al.

“A cul­ture war is a war,” he said. “There are casu­al­ties in war. And that’s what you’re see­ing.”
...

Of course, the Trump White House and reelec­tion cam­paign is claim­ing it has noth­ing to do it. So if any jour­nal­ist point out the clear con­nec­tions between this oper­a­tion and the Trump White House they will pre­sum­ably become tar­gets:

...
The infor­ma­tion unearthed by the oper­a­tion has been com­ment­ed on and spread by offi­cials inside the Trump admin­is­tra­tion and re-elec­tion cam­paign, as well as con­ser­v­a­tive activists and right-wing news out­lets such as Bre­it­bart News. In the case of the Times edi­tor, the news was first pub­lished by Bre­it­bart, imme­di­ate­ly ampli­fied on Twit­ter by Don­ald Trump Jr. and, among oth­ers, Kat­ri­na Pier­son, a senior advis­er to the Trump cam­paign, and quick­ly became the sub­ject of a Bre­it­bart inter­view with Stephanie Grisham, the White House press sec­re­tary and com­mu­ni­ca­tions direc­tor.

The White House press office said that nei­ther the pres­i­dent nor any­one in the White House was involved in or aware of the oper­a­tion, and that nei­ther the White House nor the Repub­li­can Nation­al Com­mit­tee was involved in fund­ing it.

The Trump cam­paign said it was unaware of, and not involved in, the effort, but sug­gest­ed that it served a wor­thy pur­pose. “We know noth­ing about this, but it’s clear that the media has a lot of work to do to clean up its own house,” said Tim Mur­taugh, the campaign’s com­mu­ni­ca­tions direc­tor.

The cam­paign is con­sis­tent with Mr. Trump’s long-run­ning effort to dele­git­imize crit­i­cal report­ing and brand the news media as an “ene­my of the peo­ple.” The pres­i­dent has relent­less­ly sought to dimin­ish the cred­i­bil­i­ty of news orga­ni­za­tions and cast them as polit­i­cal­ly moti­vat­ed oppo­nents.

Jour­nal­ism, he said in a tweet last week, is “noth­ing more than an evil pro­pa­gan­da machine for the Demo­c­rat Par­ty.”

The oper­a­tion has com­piled social media posts from Twit­ter, Face­book and Insta­gram, and stored images of the posts that can be pub­li­cized even if the user deletes them, said the peo­ple famil­iar with the effort. One claimed that the oper­a­tion had unearthed poten­tial­ly “fire­able” infor­ma­tion on “sev­er­al hun­dred” peo­ple.

“I am sure there will be more scalps,” said Sam Nun­berg, a for­mer aide to Mr. Trump who is a friend of Mr. Schwartz.
...

And as the fol­low­ing arti­cle describes, Arthur Schwartz has decid­ed to make this intim­i­da­tion cam­paign even more overt­ly intim­i­dat­ing by now open­ly fundrais­ing for this effort. He wants to raise at least $2 mil­lion to fund this oper­a­tion (and clear­ly wants the pub­lic to know this):

Axios

Scoop: Trump allies raise mon­ey to tar­get reporters

Mike Allen
Sep 3, 2019

Pres­i­dent Trump’s polit­i­cal allies are try­ing to raise at least $2 mil­lion to inves­ti­gate reporters and edi­tors of the New York Times, Wash­ing­ton Post and oth­er out­lets, accord­ing to a 3‑page fundrais­ing pitch reviewed by Axios.

Why it mat­ters: Trump’s war on the media is expand­ing. This group will tar­get reporters and edi­tors, while oth­er GOP 2020 enti­ties go after the social media plat­forms, alleg­ing bias, offi­cials tell us.

* The group claims it will slip dam­ag­ing infor­ma­tion about reporters and edi­tors to “friend­ly media out­lets,” such as Bre­it­bart, and tra­di­tion­al media, if pos­si­ble.
* Peo­ple involved in rais­ing the funds include GOP con­sul­tant Arthur Schwartz and the “loose net­work” that the NY Times report­ed last week is tar­get­ing jour­nal­ists. The oper­a­tions are to be run by undis­closed oth­ers.
* The prospec­tus for the new project says it’s “tar­get­ing the peo­ple pro­duc­ing the news.”

The irony: The New York Times exposed an extreme­ly impro­vi­sa­tion­al effort that had out­ed a Times edi­tor for past anti-Semit­ic tweets. This new group is now using the expo­sure to try to for­mal­ize and fund the oper­a­tion.

...

Under “Pri­ma­ry Tar­gets,” the pitch lists:

* “CNN, MSNBC, all broad­cast net­works, NY Times, Wash­ing­ton Post, Buz­zFeed, Huff­in­g­ton Post, and all oth­ers that rou­tine­ly incor­po­rate bias and mis­in­for­ma­tion in to their cov­er­age. We will also track the reporters and edi­tors of these orga­ni­za­tions.”

This isn’t an entire­ly new con­cept. The lib­er­al group Media Mat­ters mon­i­tors jour­nal­ists and pub­li­ca­tions and goes pub­lic with com­plaints of bias. But being this bla­tant and spe­cif­ic about try­ing to dis­cred­it indi­vid­ual reporters is new.

———-

“Scoop: Trump allies raise mon­ey to tar­get reporters” by Mike Allen; Axios; 09/03/2019

“CNN, MSNBC, all broad­cast net­works, NY Times, Wash­ing­ton Post, Buz­zFeed, Huff­in­g­ton Post, and all oth­ers that rou­tine­ly incor­po­rate bias and mis­in­for­ma­tion in to their cov­er­age. We will also track the reporters and edi­tors of these orga­ni­za­tions.”

Intim­i­dat­ing all of the media that does­n’t rou­tine­ly fête Trump isn’t going to be cheap. But Arthur Schwartz is pub­licly sig­nal­ing that his intim­i­da­tion oper­a­tion is going to all the resources it needs. And don’t for­get that in the age of Big Data and Cam­bridge Ana­lyt­i­ca-style mass data-col­lec­tion oper­a­tions, a lot of this oppo­si­tion research will prob­a­bly be high­ly automat­able. So if you assume that you’re too insignif­i­cant to end up being tar­get­ed by this oper­a­tion that’s prob­a­bly not a safe assump­tion. And giv­en that it’s not just jour­nal­ists, but lib­er­al activists and any­one else who open­ly oppos­es Trump (nev­er-Trumpers) that are being tar­get­ed too, it points towards the next phase of the far right’s assault on democ­ra­cy and civ­il soci­ety: micro-tar­get­ed intim­i­da­tion cam­paigns against polit­i­cal dis­si­dents. Today it’s jour­nal­ists and lib­er­al activists who don’t sup­port Trump. But in the era of social media and vast data­bas­es of bil­lions of tweets and social media posts there’s no rea­son the intim­i­da­tion needs to be lim­it­ed to jour­nal­ists or activists. Vir­tu­al­ly all cit­i­zens will poten­tial­ly be vul­ner­a­ble.

So let’s hope today’s teenagers get the memo about their social media use: watch what you post, kids, because some day it might be used against you. Espe­cial­ly by the Repub­li­can Par­ty.

There was a recent sto­ry in Politi­co that appears to solve the mys­tery of who was behind the “stringray” devices found in Wash­ing­ton DC in recent years. The exis­tence of the devices — which col­lects cell-phone data by mim­ic legit­i­mate cell-phone tow­ers — near the White House and oth­er sen­si­tive areas in DC was first pub­licly acknowl­edged by the US gov­ern­ment in April of 2018. These reports were deemed at the time to be extra alarm­ing giv­en the fact that Pres­i­dent Trump was known to use inse­cure cell­phone for sen­si­tive com­mu­ni­ca­tions. Accord­ing to the new Politi­co report, the US gov­ern­ment has con­clud­ed that the sting-ray devices were most like­ly put in place by Israel, and yet there have been no con­se­quences at all fol­low­ing this find­ing. Israel has denied the reports and Trump him­self told Politi­co, “I don’t think the Israelis were spy­ing on us...My rela­tion­ship with Israel has been great...Anything is pos­si­ble but I don’t believe it.”.

So we have reports about a US gov­ern­ment inves­ti­ga­tion con­clud­ing Israel we behind one of the most mys­te­ri­ous, and poten­tial­ly sig­nif­i­cant, spy­ing oper­a­tion uncov­ered in DC in recent years cou­pled with US gov­ern­ment denials that this hap­pened. Which is large­ly what we should have expect­ed giv­en this find­ing. On the one hand, giv­en the extreme­ly close and long-stand­ing ties between US and Israeli mil­i­tary and intel­li­gence, if this real­ly was an oper­a­tion that Israel was gen­uine­ly behind with­out the tac­it approval of the US gov­ern­ment there would like­ly be an attempt to min­i­mize the diplo­mat­ic fall­out and deal with these things qui­et­ly and out of the pub­lic eye. On the oth­er hand, if this was the kind of oper­a­tion done with the US gov­ern­men­t’s tac­it approval, we would expect at least down­play­ing of the scan­dal too.

But as the fol­low­ing arti­cle makes clear, there’s anoth­er huge we should expect the down­play­ing by the US gov­ern­ment about a sto­ry like this: The US and Israel have been increas­ing­ly out­sourc­ing their cyber-spy­ing capa­bil­i­ties to the pri­vate sec­tor and joint­ly invest­ing in these com­pa­nies. Beyond that, Jef­frey Epstein appears to be one of the fig­ures who appears to have been work­ing on this merg­ing of US and Israeli cyber-spy­ing tech­nol­o­gy in recent years. So when we talk about Israel spy­ing oper­a­tions in the US involv­ing the covert use of tech­nol­o­gy, we have to ask whether or not this was an oper­a­tion involv­ing a com­pa­ny with US nation­al secu­ri­ty ties.

The fol­low­ing report, the lat­est for Whit­ney Webb at Mint­Press
on the Epstein scan­dal, describes this grow­ing joint US/Israeli invest­ment in cyber sec­tor in recent years and some of the fig­ures behind it in addi­tion to Epstein. The piece focus­es on Car­byne (Carbyne911), the Israeli com­pa­ny start­ed in 2014 by for­mer mem­bers of Israel’s Unit 8200 cyber team. Car­byne cre­at­ed Reporty, a smart­phone app that promis­es to pro­vide faster and bet­ter com­mu­ni­ca­tions to pub­lic emer­gency first respon­ders. As we’ve seen, Reporty isn’t just a smart­phone app. It also appears to work by mon­i­tor­ing pub­lic emer­gency com­mu­ni­ca­tion sys­tems and nation­al civil­ian com­mu­ni­ca­tions infra­struc­ture for the osten­si­ble pur­pose of ensur­ing min­i­mal data loss dur­ing emer­gency response calls, which is the kind of capa­bil­i­ty with obvi­ous dual use poten­tial.

As we also saw, while for­mer Israeli prime min­is­ter Ehud Barack was pub­licly the big investor who helped start Car­byne back in 2014, it turns out Jef­frey Epstein was qui­et­ly the per­son behind Barack­’s financ­ing. Barack was a known asso­ciate of Epstein and report­ed­ly fre­quent­ed Epstein’s Man­hat­tan man­sion. So we have Epstein, a fig­ure with clear ties to Israeli intel­li­gence but also very clear ties to US intel­li­gence, invest­ing in Car­byne. Well, as the piece describes, it turns out that one of the oth­er investors in Car­byne is Peter Thiel. And Car­byne’s board of advi­sors includes for­mer Palan­tir employ­ee Trae Stephens, who was a mem­ber of the Trump tran­si­tion team. For­mer Sec­re­tary of Home­land Secu­ri­ty Michael Chertoff is also an advi­so­ry board mem­ber. These are the kinds of investors and advi­sors that make it clear Car­byne isn’t sim­ply an Israeli intel­li­gence front. This is, at a min­i­mum, a joint oper­a­tion between the US and Israel.

It’s also note­wor­thy that both Thiel and Epstein appear to have been lead­ing financiers for ‘tran­shu­man­ist’ projects like longevi­ty and arti­fi­cial intel­li­gence. Both have a his­to­ry of spon­sor­ing sci­en­tists work­ing in these areas. Both appeared to have very sim­i­lar inter­ests and moved in the same cir­cles and yet there pre­vi­ous­ly weren’t indi­ca­tions that Thiel and Epstein had a rela­tion­ship. Their mutu­al invest­ments in Car­byne helps answer that. The two def­i­nite­ly knew each oth­er because they were secret busi­ness part­ners.

How many oth­er secret busi­ness part­ner­ships might Epstein and Thiel have been involved in and now many of them involve the Israeli tech sec­tor? We obvi­ous­ly don’t know, but as the fol­low­ing arti­cle points out, Palan­tir opened an R&D branch in Israel in 2013 and there have long been sus­pi­cions that Palan­tir’s ‘pre-cog’ pre­dic­tive crime algo­rithms have been used against Pales­tin­ian pop­u­la­tions. So Palan­tir appears to be well posi­tioned to help lead any qui­et joint US-Israeli efforts to devel­op cyber-intel­li­gence capa­bil­i­ties in the pri­vate sec­tor.

Omi­nous­ly, as the arti­cle also describes, the idea of a joint US-Israeli project on ‘pre-crime’ detec­tion is one that goes back to 1982 when the “Main Core” data­base of 8 mil­lion Amer­i­cans deemed to be poten­tial sub­ver­sives was devel­oped by Oliv­er North under the “Con­ti­nu­ity of Gov­ern­ment” pro­gram and main­tained using the PROMIS soft­ware (which sounds like a com­pli­men­ta­ry pro­gram to “Rex 84”). Accord­ing to anony­mous intel­li­gence sources talk­ing to Mint­Press, this “Main Core” data­base of US cit­i­zens con­sid­ered “dis­si­dents” still exists today. Accord­ing to these anony­mous U.S. intel­li­gence offi­cials who report­ed­ly have direct knowl­edge of the US intel­li­gence community’s use of PROMIS and Main Core from the 1980s to 2000s, Israeli intel­li­gence played a role in the deploy­ment of PROMIS as the soft­ware used for the Main Core. And Palan­tir, with its PROMIS-like Inves­tiga­tive Case Man­age­ment (ICM) soft­ware already being offered to the US gov­ern­ment for use in track­ing immi­grants, is the com­pa­ny well posi­tioned to be main­tain­ing the cur­rent ver­sion of Main Core. The arti­cle also reports that Main Core was used by at least one for­mer CIA offi­cial on Ronald Reagan’s Nation­al Secu­ri­ty Coun­cil to black­mail mem­bers of Con­gress, Con­gres­sion­al staffers and jour­nal­ists. That obvi­ous­ly has the­mat­ic ties to the Epstein sex­u­al traf­fick­ing net­work that appears to have black­mail­ing pow­er­ful peo­ple as one of its core func­tions.

Also note­wor­thy in all this is is that Car­byne’s prod­ucts were ini­tial­ly sold as a solu­tion for mass shoot­ings (‘solu­tion’, in the sense that vic­tims would be able to con­tact emer­gency respon­ders). That’s part of what makes Thiel’s invest­ment in Car­byne extra inter­est­ing giv­en the pre-crime pre­dic­tion tech­nolo­gies capa­bil­i­ties Palan­tir has been offer­ing law enforce­ment in recent years. As the arti­cle notes, this all poten­tial­ly ties in to the recent push by the Trump admin­is­tra­tion to cre­ate HARPA, a new US gov­ern­ment agency mod­eled after DARPA, that could cre­ate tools for track­ing the men­tal­ly ill using smart­phones and smart­watch­es and pre­dict­ing when they might become vio­lent. Palan­tir is per­fect­ly sit­u­at­ed to cap­i­tal­ize on an ini­tia­tive like that.

And that’s all part of the con­text we have to keep in mind when read­ing reports about “string-ray” devices in Wash­ing­ton DC being set up by Israel and the response from the US gov­ern­ment is a big *yawn*. When fig­ures like Thiel and Epstein are act­ing as mid­dle-men in some sort of joint US-Israeli cyber-spy­ing pri­va­ti­za­tion dri­ve, it’s hard not to won­der if those stingray devices aren’t also part of some sort of joint ini­tia­tive:

Mint­Press

How the CIA, Mossad and “the Epstein Net­work” are Exploit­ing Mass Shoot­ings to Cre­ate an Orwellian Night­mare

Fol­low­ing anoth­er cat­a­stroph­ic mass shoot­ing or cri­sis event, Orwellian “solu­tions” are set to be foist­ed on a fright­ened Amer­i­can pub­lic by the very net­work con­nect­ed, not only to Jef­frey Epstein, but to a litany of crimes and a fright­en­ing his­to­ry of plans to crush inter­nal dis­sent in the Unit­ed States.
by Whit­ney Webb

Sep­tem­ber 06th, 2019

Fol­low­ing the arrest and sub­se­quent death in prison of alleged child sex traf­fick­er Jef­frey Epstein, a lit­tle-known Israeli tech com­pa­ny began to receive increased pub­lic­i­ty, but for all the wrong rea­sons. Not long after Epstein’s arrest, and his rela­tion­ships and finances came under scruti­ny, it was revealed that the Israeli com­pa­ny Carbyne911 had received sub­stan­tial fund­ing from Jef­frey Epstein as well as Epstein’s close asso­ciate and for­mer Prime Min­is­ter of Israel Ehud Barak, and Sil­i­con Val­ley ven­ture cap­i­tal­ist and promi­nent Trump backer Peter Thiel.

Carbyne911, or sim­ply Car­byne, devel­ops call-han­dling and iden­ti­fi­ca­tion capa­bil­i­ties for emer­gency response ser­vices in coun­tries around the world, includ­ing the Unit­ed States, where it has already been imple­ment­ed in sev­er­al U.S. coun­ties and has part­nered with major U.S. tech com­pa­nies like Google. It specif­i­cal­ly mar­kets its prod­uct as a way of mit­i­gat­ing mass shoot­ings in the Unit­ed States with­out hav­ing to change exist­ing U.S. gun laws.

Yet, Car­byne is no ordi­nary tech com­pa­ny, as it is deeply con­nect­ed to the elite Israeli mil­i­tary intel­li­gence divi­sion, Unit 8200, whose “alum­ni” often go on to cre­ate tech com­pa­nies — Car­byne among them — that fre­quent­ly main­tain their ties to Israeli intel­li­gence and, accord­ing to Israeli media reports and for­mer employ­ees, often “blur the line” between their ser­vice to Israel’s defense/intelligence appa­ra­tus and their com­mer­cial activ­i­ty. As this report will reveal, Car­byne is but one of sev­er­al Israeli tech com­pa­nies mar­ket­ing them­selves as a tech­no­log­i­cal solu­tion to mass shoot­ings that has direct ties to Israeli intel­li­gence agen­cies.

In each case, these com­pa­nies’ prod­ucts are built in such a way that they can eas­i­ly be used to ille­gal­ly sur­veil the gov­ern­ments, insti­tu­tions and civil­ians that use them, a trou­bling fact giv­en Unit 8200’s doc­u­ment­ed prowess in sur­veil­lance as a means of obtain­ing black­mail and Israel’s his­to­ry of using tech com­pa­nies to aggres­sive­ly spy on the U.S. gov­ern­ment. This is fur­ther com­pound­ed by the fact that Unit 8200-linked tech com­pa­nies have pre­vi­ous­ly received U.S. gov­ern­ment con­tracts to place “back­doors” into the U.S.’ entire telecom­mu­ni­ca­tions sys­tem as well as into the pop­u­lar prod­ucts of major Amer­i­can tech com­pa­nies includ­ing Google, Microsoft and Face­book, many of whose key man­agers and exec­u­tives are now for­mer Unit 8200 offi­cers.

...

Anoth­er fun­der of Car­byne, Peter Thiel, has his own com­pa­ny that, like Car­byne, is set to prof­it from the Trump administration’s pro­posed hi-tech solu­tions to mass shoot­ings. Indeed, after the recent shoot­ing in El Paso, Texas, Pres­i­dent Trump — who received polit­i­cal dona­tions from and has been advised by Thiel fol­low­ing his elec­tion — asked tech com­pa­nies to “detect mass shoot­ers before they strike,” a ser­vice already per­fect­ed by Thiel’s com­pa­ny Palan­tir, which has devel­oped “pre-crime soft­ware” already in use through­out the coun­try. Palan­tir is also a con­trac­tor for the U.S. intel­li­gence com­mu­ni­ty and also has a branch based in Israel.

Per­haps most dis­turb­ing of all, what­ev­er tech­no­log­i­cal solu­tion is adopt­ed by the Trump admin­is­tra­tion, it is set to use a con­tro­ver­sial data­base first devel­oped as part of a secre­tive U.S. gov­ern­ment pro­gram that involved noto­ri­ous Iran-Con­tra fig­ures like Oliv­er North as a means of track­ing and flag­ging poten­tial Amer­i­can dis­si­dents for increased sur­veil­lance and deten­tion in the event of a vague­ly defined “nation­al emer­gency.”

As this report will reveal, this data­base — often referred to as “Main Core” — was cre­at­ed with the involve­ment of Israeli intel­li­gence and Israel remained involved years after it was devel­oped, and poten­tial­ly to the present. It was also used by at least one for­mer CIA offi­cial on Pres­i­dent Reagan’s Nation­al Secu­ri­ty Coun­cil to black­mail mem­bers of Con­gress, Con­gres­sion­al staffers and jour­nal­ists, among oth­ers.

...

Demys­ti­fy­ing Car­byne

Carbyne911, which will be referred to sim­ply as Car­byne in this report, is an Israeli tech-start­up that promis­es to rev­o­lu­tion­ize how calls are han­dled by emer­gency ser­vice providers, as well as by gov­ern­ments, cor­po­ra­tions and edu­ca­tion­al insti­tu­tions. Not long after it was found­ed in 2014 by vet­er­ans of Israeli mil­i­tary intel­li­gence, Car­byne began to be specif­i­cal­ly mar­ket­ed as a solu­tion to mass shoot­ings in the Unit­ed States that goes “beyond the gun debate” and improves the “intel­li­gence that armed emer­gency respon­ders receive before enter­ing an armed shoot­er sit­u­a­tion” by pro­vid­ing video-stream­ing and acoustic input from civil­ian smart­phones and oth­er devices con­nect­ed to the Car­byne net­work.

...

As a result of increased scruti­ny of Epstein’s busi­ness activ­i­ties and his ties to Israel, par­tic­u­lar­ly to Barak, Epstein’s con­nec­tion to Car­byne was revealed and exten­sive­ly report­ed on by the inde­pen­dent media out­let Nar­a­tiv, whose exposé on Car­byne revealed not only some of the key intel­li­gence con­nec­tions of the start-up com­pa­ny but also how the archi­tec­ture of Carbyne’s prod­uct itself rais­es “seri­ous pri­va­cy con­cerns.”

Mint­Press detailed many of Carbyne’s main intel­li­gence con­nec­tions in Part III of the inves­tiga­tive series “Inside the Jef­frey Epstein Scan­dal: Too Big to Fail.” In addi­tion to Barak — for­mer Israeli prime min­is­ter and for­mer head of Israeli mil­i­tary intel­li­gence — serv­ing as Carbyne’s chair­man and a key financer, the company’s exec­u­tive team are all for­mer mem­bers of Israeli intel­li­gence, includ­ing the elite mil­i­tary intel­li­gence unit, Unit 8200, which is often com­pared to the U.S. Nation­al Secu­ri­ty Agency (NSA).

...

Notably, the NSA and Unit 8200 have col­lab­o­rat­ed on numer­ous projects, most infa­mous­ly on the Stuxnet virus as well as the Duqu mal­ware. In addi­tion, the NSA is known to work with vet­er­ans of Unit 8200 in the pri­vate sec­tor, such as when the NSA hired two Israeli com­pa­nies, to cre­ate back­doors into all the major U.S. telecom­mu­ni­ca­tions sys­tems and major tech com­pa­nies, includ­ing Face­book, Microsoft and Google. Both of those com­pa­nies, Verint and Narus, have top exec­u­tives with ties to Israeli intel­li­gence and one of those com­pa­nies, Verint (for­mer­ly Com­verse Infos­ys), has a his­to­ry of aggres­sive­ly spy­ing on U.S. gov­ern­ment facil­i­ties. Unit 8200 is also known for spy­ing on civil­ians in the occu­pied Pales­tin­ian ter­ri­to­ries for “coer­cion pur­pos­es” — i.e., gath­er­ing info for black­mail — and also for spy­ing on Pales­tin­ian-Amer­i­cans via an intel­li­gence-shar­ing agree­ment with the NSA.

Unlike many oth­er Unit 8200-linked start-ups, Car­byne also boasts sev­er­al tie-ins to the Trump admin­is­tra­tion, includ­ing Palan­tir founder and Trump ally Peter Thiel — anoth­er investor in Car­byne. In addi­tion, Carbyne’s board of advis­ers includes for­mer Palan­tir employ­ee Trae Stephens, who was a mem­ber of the Trump tran­si­tion team, as well as for­mer Sec­re­tary of Home­land Secu­ri­ty Michael Chertoff. Trump donor and New York real-estate devel­op­er Eliot Taw­ill is also on Carbyne’s board, along­side Ehud Barak and Pin­chas Buchris.

Yet, pri­va­cy con­cerns with Car­byne go beyond the company’s ties to Israeli intel­li­gence and U.S. intel­li­gence con­trac­tors like Peter Thiel. For instance, Carbyne’s smart­phone app extracts the fol­low­ing infor­ma­tion from the phones on which it is installed:

Device loca­tion, video live-streamed from the smart­phone to the call cen­ter, text mes­sages in a two-way chat win­dow, any data from a user’s phone if they have the Car­byne app and ESInet, and any infor­ma­tion that comes over a data link, which Car­byne opens in case the caller’s voice link drops out.” (empha­sis added)

...

Anoth­er cause for con­cern is how oth­er coun­tries have used plat­forms like Car­byne, which were first mar­ket­ed as emer­gency response tools, for the pur­pose of mass sur­veil­lance. Nar­a­tiv not­ed the fol­low­ing in its inves­ti­ga­tion of Car­byne:

In May, Human Rights Watch revealed Chi­nese author­i­ties use a plat­form not unlike Car­byne to ille­gal­ly sur­veil Uyghurs. China’s Inte­grat­ed Joint Oper­a­tions Plat­form brings in a much big­ger data-set and sources of video, which includes an app on people’s phones. Like Car­byne, the plat­form was designed to report emer­gen­cies. Chi­nese author­i­ties have turned it into a tool of mass sur­veil­lance.

Human Rights Watch reverse-engi­neered the app. The group dis­cov­ered the app auto­mat­i­cal­ly pro­files a user under 36 “per­son types” includ­ing “fol­low­ers of Six Lines” which is the term used to iden­ti­fy Uyghurs. Anoth­er term refers to “Hajj,” the annu­al Islam­ic pil­grim­age to Mec­ca. The app mon­i­tors every aspect of a user’s life, includ­ing per­son­al con­ver­sa­tions [and] pow­er usage, and tracks a user’s move­ment.”

Such tech­nol­o­gy is cur­rent­ly used by Israeli mil­i­tary intel­li­gence and Israel’s domes­tic intel­li­gence agency Shin Bet to jus­ti­fy “pre-crime” deten­tions of Pales­tini­ans in the occu­pied West Bank. As will be not­ed in greater detail lat­er in this report, Pales­tini­ans’ com­ments on social media are tracked by arti­fi­cial intel­li­gence algo­rithms that flag them for indef­i­nite deten­tion if they write social media posts that con­tain “trip­wire” phras­es such as “the sword of Allah.”

Carbyne’s plat­form has its own “pre-crime” ele­ments, such as it’s c‑Records com­po­nent, which stores and ana­lyzes infor­ma­tion on past calls and events that pass through its net­work. This infor­ma­tion “enables deci­sion mak­ers to accu­rate­ly ana­lyze the past and present behav­ior of their callers, react accord­ing­ly, and in time pre­dict future pat­terns.” (empha­sis added)

...

Israeli intel­li­gence, Black­mail and Sil­i­con Val­ley

Though many of the indi­vid­u­als involved in fund­ing or man­ag­ing Car­byne have proven ties to intel­li­gence, a clos­er look into sev­er­al of these play­ers reveals even deep­er con­nec­tions to both Israeli and U.S. intel­li­gence.

One of Carbyne’s clear­est con­nec­tions to Israeli intel­li­gence is through its chair­man and one of its fun­ders, Ehud Barak. Though Barak is best known for being a for­mer prime min­is­ter of Israel, he is also a for­mer min­is­ter of defense and the for­mer head of Israeli mil­i­tary intel­li­gence. He over­saw Unit 8200’s oper­a­tions, as well as oth­er units of Israeli mil­i­tary intel­li­gence, in all three of those posi­tions. For most of his mil­i­tary and lat­er polit­i­cal career, Barak has been close­ly asso­ci­at­ed with covert oper­a­tions.

...

Yet, more recent­ly, it has been Barak’s close rela­tion­ship to Epstein that has raised eye­brows and opened him up to polit­i­cal attacks from his rivals. Epstein and Barak were first intro­duced by for­mer Israeli Prime Min­is­ter Shi­mon Peres in 2002, a time when Epstein’s pedophile black­mail and sex traf­fick­ing oper­a­tion was in full swing.

...

In 2015, Barak formed a lim­it­ed part­ner­ship com­pa­ny in Israel for the explic­it pur­pose of invest­ing in Car­byne (then known as Reporty) and invest­ed mil­lions of dol­lars in the com­pa­ny, quick­ly becom­ing a major share­hold­er and sub­se­quent­ly the company’s pub­lic face and the chair­man of its board. At least $1 mil­lion of the mon­ey invest­ed in this Barak-cre­at­ed com­pa­ny that was lat­er used to invest in Car­byne came from the South­ern Trust Com­pa­ny, which was owned by Jef­frey Epstein.

In July, Bloomberg report­ed that Epstein’s South­ern Trust Com­pa­ny is iden­ti­fied in U.S. Vir­gin Islands fil­ings as “a DNA data­base and data min­ing” com­pa­ny. Giv­en Carbyne’s clear poten­tial for data-min­ing and civil­ian pro­fil­ing, Epstein’s invest­ment in Car­byne using this spe­cif­ic com­pa­ny sug­gests that Carbyne’s investors have long been aware of this lit­tle adver­tised aspect of Carbyne’s prod­uct.

In a state­ment to the Israeli news­pa­per Haaretz, Barak assert­ed:

I saw the busi­ness oppor­tu­ni­ty and reg­is­tered a part­ner­ship in my con­trol in Israel. A small num­ber of peo­ple I know invest in it…Since these are pri­vate invest­ments, it wouldn’t be prop­er or right for me to expose the investors’ details.”

How­ev­er, Barak lat­er admit­ted that Epstein had been one of the investors.

Mint­Press’ recent series on the Jef­frey Epstein scan­dal not­ed in detail Epstein’s ties to CIA/Mossad intel­li­gence assets, such as Adnan Khashog­gi; CIA front com­pa­nies, such as South­ern Air Trans­port; and orga­nized crime, through his close asso­ci­a­tion with Leslie Wexn­er. In addi­tion, Epstein’s long-time “girl­friend” and alleged madam, Ghis­laine Maxwell, has fam­i­ly links to Israeli intel­li­gence through her father, Robert Maxwell. While it appears that Epstein may have been work­ing for more than one intel­li­gence agency, Zev Shalev, for­mer exec­u­tive pro­duc­er for CBS News and jour­nal­ist at Nar­a­tiv, recent­ly stat­ed that he had inde­pen­dent­ly con­firmed with two uncon­nect­ed sources “close­ly con­nect­ed to the Epstein sto­ry and in a posi­tion to know” that Epstein had “worked for Israeli mil­i­tary intel­li­gence.”

Exclu­sive: We have two inde­pen­dent sources con­firm­ing Jef­frey Epstein worked for Israeli mil­i­tary intel­li­gence. In each case the source is close­ly con­nect­ed to the Epstein sto­ry and in a posi­tion to know. You can take it to the bank. @narativlive https://t.co/BdK1DrZEO6

— Zev Shalev (@ZevShalev) August 20, 2019

Notably, Epstein, who was known for his inter­est in obtain­ing black­mail through the sex­u­al abuse of the under­aged girls he exploit­ed, also claimed to have “dam­ag­ing infor­ma­tion” on promi­nent fig­ures in Sil­i­con Val­ley. In a con­ver­sa­tion last year with New York Times reporter James Stew­art, Epstein claimed to have “poten­tial­ly dam­ag­ing or embar­rass­ing” infor­ma­tion on Sil­i­con Valley’s elite and told Stew­art that these top fig­ures in the Amer­i­can tech indus­try “were hedo­nis­tic and reg­u­lar users of recre­ation­al drugs.” Epstein also told Stew­art that he had “wit­nessed promi­nent tech fig­ures tak­ing drugs and arrang­ing for sex” and claimed to know “details about their sup­posed sex­u­al pro­cliv­i­ties.”

In the lead-up to his recent arrest, Jef­frey Epstein appeared to have been attempt­ing to rebrand as a “tech investor,” as he had done inter­views with sev­er­al jour­nal­ists includ­ing Stew­art about tech­nol­o­gy invest­ing in the months before he was hit with fed­er­al sex traf­fick­ing charges.

...

It is unknown whether Epstein’s “dam­ag­ing infor­ma­tion” and appar­ent black­mail on notable indi­vid­u­als in the Amer­i­can tech­nol­o­gy indus­try were used to advance the objec­tives of Car­byne, which recent­ly part­nered with tech giants Google and Cis­co Sys­tems — and, more broad­ly, the expan­sion of Israeli intel­li­gence-linked tech com­pa­nies into the Amer­i­can tech sec­tor, par­tic­u­lar­ly through the acqui­si­tion of Israeli tech start-ups linked to Unit 8200 by major U.S. tech com­pa­nies.

...

Carbyne’s ties to U.S. intel­li­gence

While Epstein and Barak are the two financiers of Car­byne whose ties to intel­li­gence are clear­est, anoth­er fun­der of Car­byne, Peter Thiel, has ties to U.S. intel­li­gence and a his­to­ry of invest­ing in oth­er com­pa­nies found­ed by for­mer mem­bers of Unit 8200. Thiel co-found­ed and still owns a con­trol­ling stake in the com­pa­ny Palan­tir, which was ini­tial­ly fund­ed with a $2 mil­lion invest­ment from the CIA’s ven­ture cap­i­tal fund In-Q-Tel and quick­ly there­after became a con­trac­tor for the CIA.

After the suc­cess of its con­tract with the CIA, Palan­tir became a con­trac­tor for a vari­ety of fed­er­al agen­cies, includ­ing the FBI, the Defense Intel­li­gence Agency (DIA), the Nation­al Secu­ri­ty Agency (NSA), the Depart­ment of Home­land Security(DHS) and the military’s Spe­cial Oper­a­tions Com­mand, among oth­ers. Last year, it won a con­tract to cre­ate a new bat­tle­field intel­li­gence sys­tem for the U.S. Army. Palan­tir is also in demand for its “pre-crime tech­nol­o­gy,” whichhas been used by sev­er­al U.S. police depart­ments. Accord­ing to the Guardian, “Palan­tir tracks every­one from poten­tial ter­ror­ist sus­pects to cor­po­rate fraud­sters, child traf­fick­ers and what they refer to as ‘sub­ver­sives’… it is all done using pre­dic­tion.”

Thiel has gained atten­tion in recent years for his sup­port of Pres­i­dent Trump and for becom­ing an advis­er to Trump fol­low­ing the 2016 elec­tion, when he was “a major force in the tran­si­tion,” accord­ing to Politi­co, and “helped fill posi­tions in the Trump admin­is­tra­tion with for­mer staff.” One of those for­mer staffers was Trae Stephens, who is also on Carbyne’s board of advis­ers. Thiel also has busi­ness ties to Trump’s son-in-law and influ­en­tial advis­er, Jared Kush­n­er, as well as to Kushner’s broth­er Josh. A senior Trump cam­paign aide told Politi­co in 2017 that “Thiel is immense­ly pow­er­ful with­in the admin­is­tra­tion through his con­nec­tion to Jared.”

...

Anoth­er Car­byne-con­nect­ed indi­vid­ual worth not­ing is the for­mer head of the Depart­ment of Home­land Secu­ri­ty, Michael Chertoff, who serves on Carbyne’s board of advis­ers. In addi­tion to Chertoff’s ties to DHS, Chertoff’s com­pa­ny, The Chertoff Group, employ­ees sev­er­al promi­nent for­mer mem­bers of the U.S. intel­li­gence com­mu­ni­ty as prin­ci­pals, includ­ing Michael Hay­den, for­mer direc­tor of the CIA and for­mer direc­tor of the NSA; and Charles Allen, for­mer assis­tant direc­tor of Cen­tral Intel­li­gence for Col­lec­tion at the CIA, who worked at the agency for over 40 years.

...

Meld­ing into Sil­i­con Val­ley

Beyond its trou­bling con­nec­tions to Sil­i­con Val­ley oli­garchs, Israeli mil­i­tary intel­li­gence and the U.S.-military indus­tri­al com­plex, Carbyne’s recent part­ner­ships with two spe­cif­ic tech­nol­o­gy com­pa­nies — Google and Cis­co Sys­tems — raise even more red flags.

Car­byne announced its part­ner­ship with Cis­co Sys­tems this past April, with the lat­ter announc­ing that it would begin “align­ing its uni­fied call man­ag­er with Carbyne’s call-han­dling plat­form, allow­ing emer­gency call cen­ters to col­lect data from both 911 callers and near­by gov­ern­ment-owned IoT [Inter­net of Things] devices.” A report on the part­ner­ship pub­lished by Gov­ern­ment Tech­nol­o­gy mag­a­zine stat­ed that “Carbyne’s plat­form will be inte­grat­ed into Cis­co Kinet­ic for Cities, an IoT data plat­form that shares data across com­mu­ni­ty infra­struc­ture, smart city solu­tions, appli­ca­tions and con­nect­ed devices.” The report also not­ed that “Car­byne will also be the only 911 solu­tion in the Cis­co Mar­ket­place.”

As part of the part­ner­ship, Carbyne’s Pres­i­dent of North Amer­i­can Oper­a­tions Paul Tatro told Gov­ern­ment Tech­nol­o­gy that the Car­byne plat­form would com­bine the data it obtains from smart­phones and oth­er Car­byne-con­nect­ed devices with “what’s avail­able through near­by Cis­co-con­nect­ed road cam­eras, road­side sen­sors, smart street­lamps, smart park­ing meters or oth­er devices.” Tatro fur­ther assert­ed that “Car­byne can also ana­lyze data that’s being col­lect­ed by Cis­co IoT devices … and alert 911 auto­mat­i­cal­ly, with­out any per­son mak­ing a phone call, if there appears to be a wor­thy prob­lem,” and expressed his view that soon most emer­gency calls will not be made by human beings but “by smart cars, telem­at­ics or oth­er smart city devices.”

A few months after part­ner­ing with Cis­co Sys­tems, Car­byne announced its part­ner­ship with Google on July 10, just three days after Car­byne fun­der Jef­frey Epstein was arrest­ed in New York on fed­er­al sex traf­fick­ing charges. Carbyne’s press release of the part­ner­ship described how the com­pa­ny and Google would be team­ing up in Mex­i­co “to offer advanced mobile loca­tion to emer­gency com­mu­ni­ca­tions cen­ters (ECCs) through­out Mex­i­co” fol­low­ing the con­clu­sion of a suc­cess­ful four-week pilot pro­gram between Car­byne and Google in the Cen­tral Amer­i­can nation.

The press release also stat­ed:

Car­byne will pro­vide Google’s Android ELS (Emer­gency Loca­tion Ser­vice) in real time from emer­gency calls made on AndroidTM devices. Deploy­ment for any ECC in the coun­try won’t require any inte­gra­tion, with Car­byne pro­vid­ing numer­ous options for con­nec­tion to their secure ELS Gate­way once an ECC is approved. The Car­byne auto­mat­ed plat­form, requir­ing no human inter­ac­tion, has the poten­tial to save thou­sands of lives each year through­out Mex­i­co.”

The rea­son Carybne’s part­ner­ships with Cis­co Sys­tems and Google are sig­nif­i­cant lies in the role that Cis­co and for­mer Google CEO Eric Schmidt have played in the cre­ation of a con­tro­ver­sial “incu­ba­tor” for Israeli tech start-ups with deep ties to Israeli mil­i­tary intel­li­gence, Amer­i­can neo­con­ser­v­a­tive donor Paul Singer, and the U.S.’ Nation­al Secu­ri­ty Agency (NSA).

This com­pa­ny, called Team8, is an Israeli com­pa­ny-cre­ation plat­form whose CEO and co-founder is Nadav Zafrir, for­mer com­man­der of Unit 8200. Two of the company’s oth­er three co-founders are also “alum­ni” of Unit 8200. Among Team8’s top investors is Schmidt, the for­mer CEO of Google, who also joined Peter Thiel in fund­ing the Unit 8200-linked Bill­Guard, as well as major tech com­pa­nies includ­ing Cis­co Sys­tems and Microsoft.

Last year, Team8 con­tro­ver­sial­ly hired the for­mer head of the NSA and U.S. Cyber Com­mand, Retired Admi­ral Mike Rogers, and Zafrir stat­ed that his inter­est in hir­ing Rogers was that Rogers would be “instru­men­tal in help­ing strate­gize” Team8’s expan­sion in the Unit­ed States. Jake Williams, a vet­er­an of NSA’s Tai­lored Access Oper­a­tions (TAO) hack­ing unit, told Cyber­Scoop:

Rogers is not being brought into this role because of his tech­ni­cal expe­ri­ence. …It’s pure­ly because of his knowl­edge of clas­si­fied oper­a­tions and his abil­i­ty to influ­ence many in the U.S. gov­ern­ment and pri­vate-sec­tor con­trac­tors.”

...

Mossad gets its own In-Q-Tel

This “delib­er­ate pol­i­cy” of Netanyahu’s also recent­ly result­ed in the cre­ation of a Mossad-run ven­ture cap­i­tal fund that is specif­i­cal­ly focused on financ­ing Israeli tech start-ups. The ven­ture cap­i­tal fund, called Lib­er­tad, was first announced by Israel’s Prime Minister’s Office and was cre­at­ed with the explic­it pur­pose of “increas­ing the Israeli intel­li­gence agency’s knowl­edge base and fos­ter­ing col­lab­o­ra­tion with Israel’s vibrant start­up scene” It was mod­eled after the CIA’s ven­ture cap­i­tal fund In-Q-Tel, which invest­ed in sev­er­al Sil­i­con Val­ley com­pa­nies turned gov­ern­ment and intel­li­gence con­trac­tors — includ­ing Google and Palan­tir — with a sim­i­lar goal in mind.

Lib­er­tad declines to reveal the recip­i­ents of its fund­ing, but announced last Decem­ber that it had cho­sen five com­pa­nies in the fields of robot­ics, ener­gy, encryp­tion, web intel­li­gence, and nat­ur­al lan­guage pro­cess­ing and text analy­sis. In regard to its inter­est in web intel­li­gence, a Mossad employ­ee told the Jerusalem Post that the intel­li­gence agency was specif­i­cal­ly inter­est­ed in “inno­v­a­tive tech­nolo­gies for [the] auto­mat­ic iden­ti­fi­ca­tion of per­son­al­i­ty char­ac­ter­is­tics – per­son­al­i­ty pro­fil­ing – based on online behav­ior and activ­i­ty, using meth­ods based on sta­tis­tics, machine learn­ing, and oth­er areas.” (empha­sis added)

...

The road to fas­cism, paved by a cor­rupt­ed PROMIS

Though Israeli intelligence’s inter­est in tech com­pa­nies goes back sev­er­al years, there is a well-doc­u­ment­ed his­to­ry of Israeli intel­li­gence using bugged soft­ware to sur­veil and gain “back­door” access to gov­ern­ment data­bas­es around the world, par­tic­u­lar­ly in the Unit­ed States.

...

While the PROMIS soft­ware is per­haps best known for offer­ing Israeli intel­li­gence a back­door into as many as 80 intel­li­gence agen­cies and oth­er sen­si­tive loca­tions around the world for near­ly a decade, it was also used for a very dif­fer­ent pur­pose by promi­nent offi­cials linked to Iran-Con­tra.

One key Iran-Con­tra fig­ure — Lt. Col. Oliv­er North, then serv­ing on the Nation­al Secu­ri­ty Coun­cil — decid­ed to use PROMIS nei­ther for espi­onage nor for for­eign pol­i­cy. Instead, North turned PROMIS’ pow­er against Amer­i­cans, par­tic­u­lar­ly per­ceived dis­si­dents, a fact that remained unknown for years.

Begin­ning in 1982, as part of the high­ly clas­si­fied Con­ti­nu­ity of Gov­ern­ment (COG) pro­gram, North used the PROMIS soft­wareat a 6,100-square-foot “com­mand cen­ter” in the Depart­ment of Jus­tice, as well as at a small­er oper­a­tions room at the White House, to com­pile a list of Amer­i­can dis­si­dents and “poten­tial trou­ble­mak­ers” if the COG pro­to­col was ever invoked.

Accord­ing to a senior gov­ern­ment offi­cial with a high-rank­ing secu­ri­ty clear­ance and ser­vice in five pres­i­den­tial admin­is­tra­tions who spoke to Radar in 2008 , this was:

A data­base of Amer­i­cans, who, often for the slight­est and most triv­ial rea­son, are con­sid­ered unfriend­ly, and who, in a time of pan­ic might be incar­cer­at­ed. The data­base can iden­ti­fy and locate per­ceived ‘ene­mies of the state’ almost instan­ta­neous­ly.”

In 1993, Wired described North’s use of PROMIS in com­pil­ing this data­base as fol­lows:

Using PROMIS, sources point out, North could have drawn up lists of any­one ever arrest­ed for a polit­i­cal protest, for exam­ple, or any­one who had ever refused to pay their tax­es. Com­pared to PROMIS, Richard Nixon’s ene­mies list or Sen. Joe McCarthy’s black­list look down­right crude.”

The COG pro­gram defined this “time of pan­ic” as “a nation­al cri­sis, such as nuclear war, vio­lent and wide­spread inter­nal dis­sent, or nation­al oppo­si­tion to a US mil­i­tary inva­sion abroad,” where­by the gov­ern­ment would sus­pend the Con­sti­tu­tion, declare mar­tial law, and incar­cer­ate per­ceived dis­si­dents and oth­er “unfriend­lies” in order to pre­vent the government’s (or then-serv­ing administration’s) over­throw.

This secre­tive data­base has often been referred to as “Main Core” by gov­ern­ment insid­ers and, most trou­bling of all, it still exists today. Jour­nal­ist Christ Ketcham, cit­ing senior gov­ern­ment offi­cials, report­ed in 2008 that, at that time, Main Core was believed to con­tain the names of as many as 8 mil­lion Amer­i­cans. Eleven years lat­er, it is high­ly like­ly that the num­ber of Amer­i­cans includ­ed in the Main Core data­base has grown con­sid­er­ably.

Author and inves­tiga­tive jour­nal­ist Tim Shorrock also cov­ered oth­er dis­turb­ing aspects of the evo­lu­tion of Main Core back in 2008 for Salon. At the time, Shorrock report­ed that the George W. Bush admin­is­tra­tion was believed to have used Main Core to guide its domes­tic sur­veil­lance activ­i­ties fol­low­ing the Sep­tem­ber 11 attacks.

Cit­ing “sev­er­al for­mer U.S. gov­ern­ment offi­cials with exten­sive knowl­edge of intel­li­gence oper­a­tions,” Shorrock fur­ther not­ed that Main Core — as it was 11 years ago at the time his report was pub­lished — was said to con­tain “a vast amount of per­son­al data on Amer­i­cans, includ­ing NSA inter­cepts of bank and cred­it card trans­ac­tions and the results of sur­veil­lance efforts by the FBI, the CIA and oth­er agen­cies.”

Bill Hamil­ton, for­mer NSA intel­li­gence offi­cer and the orig­i­nal cre­ator of the PROMIS soft­ware, told Shorrock at the time that he believed that “U.S. intel­li­gence uses PROMIS as the pri­ma­ry soft­ware for search­ing the Main Core data­base” and had been told as much by an intel­li­gence offi­cial in 1992 and an NSA offi­cial in 1995. Dan Mur­phy, for­mer deputy direc­tor at the CIA, had told Hamil­ton that the NSA’s use of PROMIS was “so seri­ous­ly wrong that mon­ey alone can­not cure the prob­lem.” “I believe in ret­ro­spect that Mur­phy was allud­ing to Main Core,” Hamil­ton had told Shorrock.

Though most report­ing on Main Core, from the time its exis­tence was first revealed to the present, has treat­ed the data­base as some­thing used by the U.S. gov­ern­ment and U.S. intel­li­gence for domes­tic pur­pos­es, Mint­Press has learned that Israeli intel­li­gence was also involved with the cre­ation of the Main Core data­base. Accord­ing to a for­mer U.S. intel­li­gence offi­cial with direct knowl­edge of the U.S. intel­li­gence community’s use of PROMIS and Main Core from the 1980s to 2000s, Israeli intel­li­gence played a role in the U.S. government’s deploy­ment of PROMIS as the soft­ware used for the Main Core domes­tic sur­veil­lance data­base sys­tem.

Israeli intel­li­gence remained involved with Main Core at the time of the August 1991 death of jour­nal­ist Dan­ny Caso­laro, who was inves­ti­gat­ing not only the government’s mis­use of the stolen PROMIS soft­ware but also the Main Core data­base. This same offi­cial, who chose to remain anony­mous, told Mint­Press that, short­ly before his death, Caso­laro had obtained copies of com­put­er print­outs from the PROMIS-based Main Core domes­tic sur­veil­lance data­base sys­tem from NSA whistle­blow­er Alan Stan­dorf, who was found mur­dered a few months before Casolaro’s life­less body would be found in a West Vir­ginia hotel room.

The source also stat­ed that Main Core’s con­tents had been used for the polit­i­cal black­mail of mem­bers of Con­gress and their staff, jour­nal­ists, and oth­ers by Wal­ter Ray­mond, a senior CIA covert oper­a­tor in psy­ops and dis­in­for­ma­tion who served on Pres­i­dent Reagan’s Nation­al Secu­ri­ty Coun­cil dur­ing and after Main Core’s cre­ation. If used for this pur­pose by Ray­mond in the 1980s, Main Core has also like­ly been used by oth­er indi­vid­u­als with access to the data­base for black­mail­ing pur­pos­es in the years since.

...

Peter Thiel’s See­ing Stone

As was men­tioned ear­li­er in this report, Palan­tir — the com­pa­ny co-found­ed by Peter Thiel — is set to prof­it hand­some­ly from the Trump administration’s plans to use its “pre-crime” tech­nol­o­gy, which is already used by police depart­ments through­out the coun­try and also used to track Amer­i­cans based on the company’s inte­gra­tive data-min­ing approach. Palan­tir, named for the “see­ing stones” in the Lord of the Rings nov­els, also mar­kets soft­ware to for­eign (and domes­tic) intel­li­gence agen­cies that pre­dicts the like­li­hood that an indi­vid­ual will com­mit an act of ter­ror­ism or vio­lence.

Aside from its “pre-crime” prod­ucts, Palan­tir has come under fire in recent years as a result of the company’s con­tracts with Immi­gra­tion and Cus­toms Enforce­ment (ICE), where it cre­at­ed an intel­li­gence sys­tem known as Inves­tiga­tive Case Man­age­ment (ICM). The IB Times described ICM as “a vast ‘ecosys­tem’ of data to help immi­gra­tion offi­cials in iden­ti­fy­ing tar­gets and cre­at­ing cas­es against them” and also “pro­vides ICE agents with access to data­bas­es man­aged by oth­er fed­er­al agen­cies.” ICM fur­ther gives ICE access to “tar­gets’ per­son­al and sen­si­tive infor­ma­tion, such as back­ground on school­ing, employ­ment, fam­i­ly rela­tion­ships, phone records, immi­gra­tion his­to­ry, bio­met­rics data, crim­i­nal records as well as home and work address­es.” In oth­er words, Palantir’s ICM is essen­tial­ly a “Main Core” for immi­grants.

Notably, part of Oliv­er North’s orig­i­nal inten­tions in “Main Core” was to track immi­grants then com­ing from Cen­tral Amer­i­ca as well as Amer­i­cans who opposed Rea­gan era pol­i­cy with respect to Cen­tral Amer­i­ca. At that time, Main Core was believed to be con­trolled by the Fed­er­al Emer­gency Man­age­ment Admin­is­tra­tion (FEMA), which is now part of the Depart­ment of Home­land Secu­ri­ty (DHS).

...

If the Trump admin­is­tra­tion moves for­ward with its pro­pos­al of employ­ing tech­nol­o­gy to detect poten­tial mass shoot­ers before they strike, Palantir’s tech­nol­o­gy is set to be used, giv­en that it has already been used by U.S. law enforce­ment and U.S. intel­li­gence to deter­mine which peo­ple run “the high­est risk of being involved in gun vio­lence,” accord­ing to an inves­ti­ga­tion of Palan­tir by The Verge. Fur­ther­more, Palantir’s close ties to the Trump admin­is­tra­tion make the company’s role in a future nation­wide “pre-crime” pre­ven­tion sys­tem based on tech­nol­o­gy appear inevitable.

Worse still is the appar­ent over­lap between Palan­tir and Main Core. Palan­tir — which has obvi­ous sim­i­lar­i­ties to PROMIS — is already known to use its soft­ware to track poten­tial ter­ror threats, includ­ing domes­tic ter­ror threats, and a cat­e­go­ry of peo­ple it refers to as “sub­ver­sives.” Palantir’s track­ing of these indi­vid­u­als “is all done using pre­dic­tion.” Palantir’s close ties to the U.S. intel­li­gence com­mu­ni­ty sug­gest that Palan­tir may already have access to the Main Core data­base. Tim Shorrock told Mint­Press that Palantir’s use of Main Core is “cer­tain­ly pos­si­ble,” par­tic­u­lar­ly in light of the company’s use of the term “sub­ver­sive” to describe a cat­e­go­ry of peo­ple that its soft­ware tracks.

Palan­tir also has alleged ties to Israeli intel­li­gence, as there have long been sus­pi­cions that Israeli intel­li­gence has used Palan­tir as part of its AI “pre-crime” algo­rithms tar­get­ing Pales­tini­ans after Palan­tir opened a research and devel­op­ment (R&D) cen­ter in Israel in 2013. The cur­rent head of Palan­tir Israel, Hamul­tal Meri­dor, pre­vi­ous­ly found­ed a brain-machine inter­face orga­ni­za­tion and was senior direc­tor of web intel­li­gence at Verint (for­mer­ly Com­verse Infos­ys), which has deep con­nec­tions to Unit 8200, a his­to­ry of espi­onage in the Unit­ed States and was one of the two com­pa­nies con­tract­ed by the NSA to insert a “back­door” into the U.S. telecom­mu­ni­ca­tions sys­tem and pop­u­lar prod­ucts of major Amer­i­can tech com­pa­nies.

Giv­en the above, Peter Thiel’s 2018 deci­sion to fund Car­byne, the Unit 8200-linked start-up that mar­kets itself as a tech­no­log­i­cal solu­tion to mass shoot­ings in the U.S., strong­ly sug­gests that Thiel has been antic­i­pat­ing for some time the now-pub­lic efforts of the Trump admin­is­tra­tion to employ “pre-crime” tech­nol­o­gy to track and tar­get Amer­i­cans who show signs of “men­tal ill­ness” and “vio­lent ten­den­cies.”

A night­mare even Orwell could not have pre­dict­ed

In ear­ly August, in the wake of the shoot­ing at an El Paso Wal­mart, Pres­i­dent Trump called on big tech com­pa­nies to col­lab­o­rate with the Jus­tice Depart­ment in the cre­ation of soft­ware that “stops mass mur­ders before they start” by detect­ing poten­tial mass shoot­ers before they cnm act. Though Trump’s ideas were short on specifics, there is now a new pro­pos­al that would cre­ate a new gov­ern­ment agency that will use data gath­ered from civil­ian elec­tron­ic devices to iden­ti­fy “neu­robe­hav­ioral” warn­ing signs, there­by flag­ging “poten­tial shoot­ers” for increased sur­veil­lance and poten­tial­ly deten­tion.

This new agency, as pro­posed by the foun­da­tion led by for­mer NBC Uni­ver­sal pres­i­dent and vice chair­man of Gen­er­al Elec­tric Robert Wright, would be known as the Health Advanced Research Projects Agency (HARPA) and would be mod­eled after the Defense Advanced Research Projects Agency (DARPA). Per the pro­pos­al, recent­ly detailed by the Wash­ing­ton Post, the flag­ship pro­gram of HARPA would be “Safe Home” (Stop­ping Aber­rant Fatal Events by Help­ing Over­come Men­tal Extremes), which would use “break­through tech­nolo­gies with high speci­fici­ty and sen­si­tiv­i­ty for ear­ly diag­no­sis of neu­ropsy­chi­atric vio­lence,” specif­i­cal­ly “advanced ana­lyt­i­cal tools based on arti­fi­cial intel­li­gence and machine learn­ing.”

The pro­gram would cost an esti­mat­ed $60 mil­lion over four years and would use data from “Apple Watch­es, Fit­bits, Ama­zon Echo and Google Home” and oth­er con­sumer elec­tron­ic devices, as well as infor­ma­tion pro­vid­ed by health-care providers to iden­ti­fy who may be a threat.

The Wash­ing­ton Post report­ed that Pres­i­dent Trump has react­ed “very pos­i­tive­ly” to the pro­pos­al and that he was “sold on the con­cept.” The Post also not­ed that Wright sees the president’s daugh­ter, Ivan­ka, as “the most effec­tive cham­pi­on of the pro­pos­al and has pre­vi­ous­ly briefed her on HARPA him­self.” Ivan­ka has pre­vi­ous­ly been cit­ed as a dri­ving force behind some of her father’s pol­i­cy deci­sions, includ­ing his deci­sion to bomb Syr­ia after an alleged chem­i­cal weapons attack in 2017.

...

For any­one famil­iar with DARPA, such claims should imme­di­ate­ly sound loud alarm bells, espe­cial­ly since DARPA is already devel­op­ing its own solu­tion to “men­tal health” issues in the form of a “brain-machine inter­face” as part of its N3 pro­gram. That pro­gram, accord­ing to reports, involves “non­in­va­sive and ‘minute­ly’ inva­sive neur­al inter­faces to both read and write into the brain,” help dis­tance sol­diers “from the emo­tion­al guilt of war­fare” by “cloud­ing their per­cep­tion” and “to pro­gram arti­fi­cial mem­o­ries of fear, desire, and expe­ri­ences direct­ly into the brain.” Though N3 is intend­ed to improve the prowess of Amer­i­can sol­diers, it is also set to be used as a means of pur­su­ing DARPA’s Sys­tems-Based Neu­rotech­nol­o­gy for Emerg­ing Ther­a­pies (SUBNETS) project, which aims to “to devel­op a tiny, implant­ed chip in the skull to treat psy­chi­atric dis­or­ders such as anx­i­ety, PTSD and major depres­sion.”

...

———-

“How the CIA, Mossad and “the Epstein Net­work” are Exploit­ing Mass Shoot­ings to Cre­ate an Orwellian Night­mare” by Whit­ney Webb; Mint­Press; 09/06/2019

“Anoth­er fun­der of Car­byne, Peter Thiel, has his own com­pa­ny that, like Car­byne, is set to prof­it from the Trump administration’s pro­posed hi-tech solu­tions to mass shoot­ings. Indeed, after the recent shoot­ing in El Paso, Texas, Pres­i­dent Trump — who received polit­i­cal dona­tions from and has been advised by Thiel fol­low­ing his elec­tion — asked tech com­pa­nies to “detect mass shoot­ers before they strike,” a ser­vice already per­fect­ed by Thiel’s com­pa­ny Palan­tir, which has devel­oped “pre-crime soft­ware” already in use through­out the coun­try. Palan­tir is also a con­trac­tor for the U.S. intel­li­gence com­mu­ni­ty and also has a branch based in Israel.”

As we can see, Peter Thiel and Jef­frey Epstein’s paths did indeed cross with their mutu­al invest­ments in Car­byne. And while we should have expect­ed their paths to cross giv­en the enor­mous over­lap between their inter­ests and activ­i­ties, this is the first con­fir­ma­tion we’ve found. It’s also a big rea­son we should­n’t assume that sto­ries about Israeli spy­ing on the US gov­ern­ment aren’t being done with the US gov­ern­men­t’s par­tic­i­pa­tion. Don’t for­get that let­ting Israel spy on US cit­i­zens and oth­ers in the DC area could be a means of the US intel­li­gence ser­vices get­ting around legal and con­sti­tu­tion­al restric­tions on domes­tic sur­veil­lance. In oth­er words, there are a some poten­tial­ly huge incen­tives for a joint US-Israeli spy­ing oper­a­tion that includes spy­ing on Amer­i­cans. Espe­cial­ly if that spy­ing allows for the black­mail­ing of US politi­cians. And based on the his­to­ry of pro­grams like the “Main Core” dis­si­dent data­base that was report­ed­ly used for black­mail­ing mem­bers of con­gress, and the sup­port­ing role Israeli intel­li­gence report­ed­ly played in set­ting “Main Core” up, we should­n’t be sur­prised by any sto­ries at all about Israel spy­ing oper­a­tions in DC. Giv­en that his­to­ry, the only thing we should be sur­prised by is if this oper­a­tion was­n’t done in coor­di­na­tion with US intel­li­gence:

...
Per­haps most dis­turb­ing of all, what­ev­er tech­no­log­i­cal solu­tion is adopt­ed by the Trump admin­is­tra­tion, it is set to use a con­tro­ver­sial data­base first devel­oped as part of a secre­tive U.S. gov­ern­ment pro­gram that involved noto­ri­ous Iran-Con­tra fig­ures like Oliv­er North as a means of track­ing and flag­ging poten­tial Amer­i­can dis­si­dents for increased sur­veil­lance and deten­tion in the event of a vague­ly defined “nation­al emer­gency.”

As this report will reveal, this data­base — often referred to as “Main Core” — was cre­at­ed with the involve­ment of Israeli intel­li­gence and Israel remained involved years after it was devel­oped, and poten­tial­ly to the present. It was also used by at least one for­mer CIA offi­cial on Pres­i­dent Reagan’s Nation­al Secu­ri­ty Coun­cil to black­mail mem­bers of Con­gress, Con­gres­sion­al staffers and jour­nal­ists, among oth­ers.

...

While the PROMIS soft­ware is per­haps best known for offer­ing Israeli intel­li­gence a back­door into as many as 80 intel­li­gence agen­cies and oth­er sen­si­tive loca­tions around the world for near­ly a decade, it was also used for a very dif­fer­ent pur­pose by promi­nent offi­cials linked to Iran-Con­tra.

One key Iran-Con­tra fig­ure — Lt. Col. Oliv­er North, then serv­ing on the Nation­al Secu­ri­ty Coun­cil — decid­ed to use PROMIS nei­ther for espi­onage nor for for­eign pol­i­cy. Instead, North turned PROMIS’ pow­er against Amer­i­cans, par­tic­u­lar­ly per­ceived dis­si­dents, a fact that remained unknown for years.

Begin­ning in 1982, as part of the high­ly clas­si­fied Con­ti­nu­ity of Gov­ern­ment (COG) pro­gram, North used the PROMIS soft­wareat a 6,100-square-foot “com­mand cen­ter” in the Depart­ment of Jus­tice, as well as at a small­er oper­a­tions room at the White House, to com­pile a list of Amer­i­can dis­si­dents and “poten­tial trou­ble­mak­ers” if the COG pro­to­col was ever invoked.

Accord­ing to a senior gov­ern­ment offi­cial with a high-rank­ing secu­ri­ty clear­ance and ser­vice in five pres­i­den­tial admin­is­tra­tions who spoke to Radar in 2008 , this was:

A data­base of Amer­i­cans, who, often for the slight­est and most triv­ial rea­son, are con­sid­ered unfriend­ly, and who, in a time of pan­ic might be incar­cer­at­ed. The data­base can iden­ti­fy and locate per­ceived ‘ene­mies of the state’ almost instan­ta­neous­ly.”

In 1993, Wired described North’s use of PROMIS in com­pil­ing this data­base as fol­lows:

Using PROMIS, sources point out, North could have drawn up lists of any­one ever arrest­ed for a polit­i­cal protest, for exam­ple, or any­one who had ever refused to pay their tax­es. Com­pared to PROMIS, Richard Nixon’s ene­mies list or Sen. Joe McCarthy’s black­list look down­right crude.”

The COG pro­gram defined this “time of pan­ic” as “a nation­al cri­sis, such as nuclear war, vio­lent and wide­spread inter­nal dis­sent, or nation­al oppo­si­tion to a US mil­i­tary inva­sion abroad,” where­by the gov­ern­ment would sus­pend the Con­sti­tu­tion, declare mar­tial law, and incar­cer­ate per­ceived dis­si­dents and oth­er “unfriend­lies” in order to pre­vent the government’s (or then-serv­ing administration’s) over­throw.

This secre­tive data­base has often been referred to as “Main Core” by gov­ern­ment insid­ers and, most trou­bling of all, it still exists today. Jour­nal­ist Christ Ketcham, cit­ing senior gov­ern­ment offi­cials, report­ed in 2008 that, at that time, Main Core was believed to con­tain the names of as many as 8 mil­lion Amer­i­cans. Eleven years lat­er, it is high­ly like­ly that the num­ber of Amer­i­cans includ­ed in the Main Core data­base has grown con­sid­er­ably.

Author and inves­tiga­tive jour­nal­ist Tim Shorrock also cov­ered oth­er dis­turb­ing aspects of the evo­lu­tion of Main Core back in 2008 for Salon. At the time, Shorrock report­ed that the George W. Bush admin­is­tra­tion was believed to have used Main Core to guide its domes­tic sur­veil­lance activ­i­ties fol­low­ing the Sep­tem­ber 11 attacks.

Cit­ing “sev­er­al for­mer U.S. gov­ern­ment offi­cials with exten­sive knowl­edge of intel­li­gence oper­a­tions,” Shorrock fur­ther not­ed that Main Core — as it was 11 years ago at the time his report was pub­lished — was said to con­tain “a vast amount of per­son­al data on Amer­i­cans, includ­ing NSA inter­cepts of bank and cred­it card trans­ac­tions and the results of sur­veil­lance efforts by the FBI, the CIA and oth­er agen­cies.”

Bill Hamil­ton, for­mer NSA intel­li­gence offi­cer and the orig­i­nal cre­ator of the PROMIS soft­ware, told Shorrock at the time that he believed that “U.S. intel­li­gence uses PROMIS as the pri­ma­ry soft­ware for search­ing the Main Core data­base” and had been told as much by an intel­li­gence offi­cial in 1992 and an NSA offi­cial in 1995. Dan Mur­phy, for­mer deputy direc­tor at the CIA, had told Hamil­ton that the NSA’s use of PROMIS was “so seri­ous­ly wrong that mon­ey alone can­not cure the prob­lem.” “I believe in ret­ro­spect that Mur­phy was allud­ing to Main Core,” Hamil­ton had told Shorrock.

Though most report­ing on Main Core, from the time its exis­tence was first revealed to the present, has treat­ed the data­base as some­thing used by the U.S. gov­ern­ment and U.S. intel­li­gence for domes­tic pur­pos­es, Mint­Press has learned that Israeli intel­li­gence was also involved with the cre­ation of the Main Core data­base. Accord­ing to a for­mer U.S. intel­li­gence offi­cial with direct knowl­edge of the U.S. intel­li­gence community’s use of PROMIS and Main Core from the 1980s to 2000s, Israeli intel­li­gence played a role in the U.S. government’s deploy­ment of PROMIS as the soft­ware used for the Main Core domes­tic sur­veil­lance data­base sys­tem.

Israeli intel­li­gence remained involved with Main Core at the time of the August 1991 death of jour­nal­ist Dan­ny Caso­laro, who was inves­ti­gat­ing not only the government’s mis­use of the stolen PROMIS soft­ware but also the Main Core data­base. This same offi­cial, who chose to remain anony­mous, told Mint­Press that, short­ly before his death, Caso­laro had obtained copies of com­put­er print­outs from the PROMIS-based Main Core domes­tic sur­veil­lance data­base sys­tem from NSA whistle­blow­er Alan Stan­dorf, who was found mur­dered a few months before Casolaro’s life­less body would be found in a West Vir­ginia hotel room.

The source also stat­ed that Main Core’s con­tents had been used for the polit­i­cal black­mail of mem­bers of Con­gress and their staff, jour­nal­ists, and oth­ers by Wal­ter Ray­mond, a senior CIA covert oper­a­tor in psy­ops and dis­in­for­ma­tion who served on Pres­i­dent Reagan’s Nation­al Secu­ri­ty Coun­cil dur­ing and after Main Core’s cre­ation. If used for this pur­pose by Ray­mond in the 1980s, Main Core has also like­ly been used by oth­er indi­vid­u­als with access to the data­base for black­mail­ing pur­pos­es in the years since.
...

So is the sto­ry about Israeli “stingrays” in DC real­ly just a sto­ry about an Israeli spy­ing oper­a­tion? Or is it a sto­ry about a joint US-Isre­ali spy­ing oper­a­tion? And if it is a joint oper­a­tion, is it part of a black­mail oper­a­tion too? Is Palan­tir involved? These are the kinds of ques­tions we have to ask now that we’ve learned that Peter Thiel and Jef­frey Epstein were qui­et co-investors in Israeli tech com­pa­nies with clear ‘dual use’ capa­bil­i­ties.

Here’s some arti­cles are worth keep­ing in mind regard­ing the ongo­ing ques­tion of who Jef­frey Epstein was coor­di­nat­ing with in his Sil­i­con Val­ley invest­ments and the peo­ple involved with reha­bil­i­ta­tion of Epstein’s rep­u­ta­tion in recent years. We’ve already seen how one of Epstein’s co-investors in Carbyne911 — the Israeli tech com­pa­ny that makes emer­gency respon­der com­mu­ni­ca­tion tech­nol­o­gy with what appears to be pos­si­ble ‘dual use’ intel­li­gence capa­bil­i­ties — is Peter Thiel. Epstein was report­ed­ly the financier behind the 2015 invest­ments in Car­byne by for­mer Israeli Prime Min­is­ter Ehud Barak. Thiel’s Founders Fund invest­ed in Car­byne in 2018. But as the fol­low­ing arti­cle describes, Epstein was get­ting intro­duced to major Sil­i­con Val­ley financiers like Thiel back in 2015. And it was appar­ent­ly Sil­i­con Val­ley investor Reid Hoff­man, a mem­ber of the ‘Pay­Pal Mafia’, who arranged for an August 2015 din­ner where Epstein was a guest along with Elon Musk, Mark Zucker­berg, and Peter Thiel.

Hoff­man has sub­se­quent­ly pub­licly apol­o­gized for invit­ing Epstein to this din­ner, say­ing in an email, “By agree­ing to par­tic­i­pate in any fundrais­ing activ­i­ty where Epstein was present, I helped to repair his rep­u­ta­tion and per­pet­u­ate injus­tice. For this, I am deeply regret­ful.” So Hoff­man acknowl­edges that this din­ner helped repair Epstein’s rep­u­ta­tion.

Hoff­man also acknowl­edges sev­er­al inter­ac­tions with Epstein that he says were for the pur­pose of fundrais­ing for MIT’s Media Lab, which has been reel­ing for the rev­e­la­tions of the exten­sive dona­tions it received from Epstein even after his 2009 child sex traf­fick­ing con­vic­tions. Hoff­man asserts that Epstein’s pres­ence at this din­ner was at the request of Joi Ito, then the head of Media Lab, for the pur­pose of fund-rais­ing for Media Lab. Giv­en that Epstein had already been donat­ing to MIT Media Lab for years, it’s unclear how Epstein’s pres­ence at the din­ner would assist in that fundrais­ing effort. Was Epstein sup­posed to con­vince Musk, Thiel, and Zucker­berg to donate too?

Recall that Hoff­man was report­ed­ly the fig­ure who financed the oper­a­tion by New Knowl­edge to run a fake ‘Russ­ian Bot’ net­work in the 2017 Alaba­ma spe­cial Sen­ate race. Also recall how, while Hoff­man’s polit­i­cal dona­tions are pri­mar­i­ly to Democ­rats, he’s also expressed some views strong­ly against the New Deal and gov­ern­ment reg­u­la­tions. If he’s a real Demo­c­rat, he’s decid­ed­ly in the ‘cor­po­rate Demo­c­rat’ wing of the par­ty.

So Hoff­man invit­ed Epstein to an August 2015 din­ner with lead­ing Sil­i­con Val­ley investors like Thiel, Zucker­berg, and Musk, appar­ent­ly at the request of the head of the MIT Media Lab to help with fundrais­ing despite Epstein hav­ing donat­ed to the lab for years. At least that’s the expla­na­tion we’re being giv­en for this August 2015 din­ner:

Busi­ness Insid­er

LinkedIn founder and Grey­lock part­ner Reid Hoff­man apol­o­gizes for his role in rehab­bing Jef­frey Epstein’s pub­lic image in 2015

* In an email to Axios on Thurs­day, LinkedIn founder and Grey­lock part­ner Reid Hoff­man apol­o­gized for his role in help­ing repair Jef­frey Epstein’s image in 2015.
* Hoff­man invit­ed Joi Ito, direc­tor of the MIT Media Lab, and Epstein to an August 2015 din­ner in Palo Alto with Elon Musk, Mark Zucker­berg, and Peter Thiel.
* Epstein had finan­cial­ly backed Ito’s Media Lab in addi­tion to per­son­al­ly help­ing fund Ito’s ven­ture cap­i­tal fund.
* In the email, Hoff­man says his inter­ac­tions with Epstein “came at the request of Joi Ito, for the pur­pos­es of fundrais­ing for the MIT Media Lab.”

Megan Hern­broth
09/13/2019

Reid Hoff­man, the founder of LinkedIn and one of Sil­i­con Val­ley’s most high-pro­file ven­ture cap­i­tal investors, apol­o­gized on Thurs­day for his role in help­ing to repair the image of con­vict­ed sex offend­er Jef­frey Epstein.

In an email to Axios, Hoff­man acknowl­edged sev­er­al inter­ac­tions with Epstein, which he said were for the pur­pose of fundrais­ing for MIT’s renown Media Lab. Hoff­man said he had been told that MIT had vet­ted and approved Epstein’s par­tic­i­pa­tion in fundrais­ing, but said his deci­sion to be involved with Epstein was nonethe­less a mis­take.

“By agree­ing to par­tic­i­pate in any fundrais­ing activ­i­ty where Epstein was present, I helped to repair his rep­u­ta­tion and per­pet­u­ate injus­tice. For this, I am deeply regret­ful,” Hoff­man said in the email.

Epstein’s ties to Sil­i­con Val­ley and to MIT have come under scruti­ny in recent weeks, fol­low­ing the financier’s arrest on sex traf­fick­ing charges and his sub­se­quent death by sui­cide.

Hoff­man invit­ed Joi Ito, direc­tor of the MIT Media Lab, and Epstein to an August 2015 din­ner in Palo Alto with Tes­la CEO Elon Musk, Face­book CEO Mark Zucker­berg, and Palan­tir founder Peter Thiel.

“My few inter­ac­tions with Jef­frey Epstein came at the request of Joi Ito, for the pur­pos­es of fundrais­ing for the MIT Media Lab. Pri­or to these inter­ac­tions, I was told by Joi that Epstein had cleared the MIT vet­ting process, which was the basis for my par­tic­i­pa­tion,” Hoff­man wrote.

In addi­tion to back­ing MIT Media Lab, Epstein also report­ed­ly helped per­son­al­ly finance Ito’s ven­ture cap­i­tal fund. Grey­lock, the ven­ture cap­i­tal firm at which Hoff­man is a part­ner, has denied that Epstein had invest­ed in any funds as a lim­it­ed part­ner. There remains the pos­si­bil­i­ty, how­ev­er, that Epstein invest­ed in Grey­lock and oth­ers through a ” fund of funds,” which does not have to dis­close its investors to ven­ture firms it backs.

Accord­ing to Axios, Hoff­man fund­ed the Media Lab’s Dis­obe­di­ence Award for “indi­vid­u­als and groups who engage in respon­si­ble, eth­i­cal dis­obe­di­ence aimed at chal­leng­ing norms, rules, or laws that sus­tain soci­ety’s injus­tices,” which last year went to lead­ers of the #MeToo move­ment.

Hoff­man’s email was made pub­lic only min­utes after a let­ter from MIT pres­i­dent L. Rafael Reif, which also blamed Ito for the uni­ver­si­ty’s over­sight of Epstein’s involve­ment. The let­ter report­ed “pre­lim­i­nary” find­ings of an inves­ti­ga­tion that was sparked by rev­e­la­tions that Epstein had fund­ed Ito’s Media Lab in addi­tion to his ven­ture cap­i­tal fund.

...

———-

“LinkedIn founder and Grey­lock part­ner Reid Hoff­man apol­o­gizes for his role in rehab­bing Jef­frey Epstein’s pub­lic image in 2015” by Megan Hern­broth; Busi­ness Insid­er; 09/13/2019

““By agree­ing to par­tic­i­pate in any fundrais­ing activ­i­ty where Epstein was present, I helped to repair his rep­u­ta­tion and per­pet­u­ate injus­tice. For this, I am deeply regret­ful,” Hoff­man said in the email.”

So the way Hoff­man is spin­ning this, he was help­ing to repair Epstein’s rep­u­ta­tion by hav­ing him present at this august 2015 meet­ing for “fundrais­ing activ­i­ties” for MIT’s Media Lab. And Epstein’s involve­ment in this fundrais­ing was done at the behest of Joi Ito:

...
Hoff­man invit­ed Joi Ito, direc­tor of the MIT Media Lab, and Epstein to an August 2015 din­ner in Palo Alto with Tes­la CEO Elon Musk, Face­book CEO Mark Zucker­berg, and Palan­tir founder Peter Thiel.

“My few inter­ac­tions with Jef­frey Epstein came at the request of Joi Ito, for the pur­pos­es of fundrais­ing for the MIT Media Lab. Pri­or to these inter­ac­tions, I was told by Joi that Epstein had cleared the MIT vet­ting process, which was the basis for my par­tic­i­pa­tion,” Hoff­man wrote.
...

But, again, Epstein has been donat­ed to the Media Lab for years. So why would he need to attend anoth­er fundrais­ing din­ner? Was Epstein mak­ing future dona­tions con­tin­gent on Media Lab some­how rehab­bing his rep­u­ta­tion? Or was he at this meet­ing to make a pitch to Musk, Zucker­berg, and Thiel for why they should donate to Media Lab too?

Note that, in addi­tion to Hoff­man fund­ing the Media Lab’s Dis­obe­di­ence Award, he also sites on Media Lab’s advi­so­ry coun­cil. So he’s more than just a donor and fundrais­er for Media Lab.

It’s also worth not­ing that, as the fol­low­ing arti­cle describes, some­one in Sil­i­con Val­ley appeared to be try­ing to assist Epstein in the pub­lic reha­bil­i­ta­tion of his rep­u­ta­tion as late as this sum­mer, after the Mia­mi Her­ald’s explo­sive report­ing on him in Decem­ber. So Epstein has some pret­ty huge mys­tery fans in Sil­i­con Val­ley:

Busi­ness Insid­er

Jef­frey Epstein was meet­ing with Sil­i­con Val­ley reporters before his arrest, ‘ram­bling’ about all the peo­ple he knew in tech

* Jef­frey Epstein met with at least three reporters, two of them for The New York Times, in the months lead­ing up to his arrest on child-sex-traf­fick­ing charges.
* The inter­views seemed to touch on Epstein’s rela­tion­ship with Sil­i­con Val­ley, sug­gest­ing that he was try­ing to reha­bil­i­tate his image and become known as a tech investor.
* Yes­ter­day, The New York Times pub­lished a year-old inter­view that Epstein gave to the colum­nist James B. Stew­art, but it has not pub­lished a sep­a­rate inter­view that the Times reporter Nel­lie Bowles con­duct­ed at Epstein’s Man­hat­tan man­sion before his arrest.
* A reporter for The Infor­ma­tion inter­viewed Epstein in June about “tech­nol­o­gy invest­ing.” The site’s edi­tor-in-chief said Epstein “ram­bled about peo­ple he knew in the indus­try” but that she was­n’t pub­lish­ing the inter­view because it “was­n’t news­wor­thy.”

John Cook
Aug. 13, 2019, 2:33 PM

The new­ly deceased sex crim­i­nal Jef­frey Epstein spoke from beyond the grave yes­ter­day, thanks to report from the New York Times colum­nist James B. Stew­art, who spilled his note­book from a year-old “back­ground” inter­view Epstein had giv­en at his Man­hat­tan man­sion.

Busi­ness Insid­er has learned that Stew­art isn’t the only reporter that vis­it­ed Epstein in recent months. The sex offend­er also grant­ed inter­views to anoth­er New York Times reporter, Nel­lie Bowles, and a reporter for tech site The Infor­ma­tion in the weeks and months lead­ing up to his most recent arrest on child-sex-traf­fick­ing charges in July. Nei­ther The Times nor The Infor­ma­tion has yet pub­lished the fruits of those inter­views, and the edi­tor-in-chief of The Infor­ma­tion said she had no plans to do so.

Epstein’s meet­ings with reporters, one of which took place as recent­ly as June, sug­gest that the dis­graced financier was try­ing to reha­bil­i­tate his image — or at least fos­ter rela­tion­ships with news out­lets — even as fed­er­al pros­e­cu­tors were clos­ing in.

All three inter­views seem to have touched on Epstein’s rela­tion­ship with Sil­i­con Val­ley. Stew­art wrote that he con­tact­ed Epstein to con­firm a rumor that Epstein was advis­ing Tes­la founder Elon Musk, and both The Infor­ma­tion and Bowles cov­er the tech sec­tor. Stew­art reached out direct­ly to Epstein, but it’s unclear who bro­kered the oth­er meet­ings. The tech focus sug­gests that some­one in Sil­i­con Val­ley may have been try­ing to help Epstein con­nect with reporters.

A jour­nal­ist for The Infor­ma­tion met with Epstein in June to dis­cuss “tech­nol­o­gy invest­ing,” Jes­si­ca Lessin, the site’s edi­tor-in-chief, con­firmed to Busi­ness Insid­er. That was just weeks before his July arrest and sev­en months after the Mia­mi Her­ald’s bru­tal inves­ti­ga­tion laid bare the extent to which Epstein escaped account­abil­i­ty for his crimes against under­age vic­tims.

“One of our reporters met with Jef­frey Epstein, in June, to talk about tech­nol­o­gy invest­ing,” Lessin said in a state­ment to Busi­ness Insid­er. “This was before his July arrest. She was intro­duced to him because he was believed to be an investor in ven­ture cap­i­tal funds, which we could not ver­i­fy. The dis­cus­sion was­n’t news­wor­thy; he ram­bled about peo­ple he knew in the indus­try. His death has not changed our judg­ment about the news­wor­thi­ness.”

Since Epstein’s arrest in July, his con­nec­tions to fig­ures in the tech, finan­cial, phil­an­thropic, polit­i­cal, and sci­en­tif­ic worlds have become of intense inter­est to reporters, who have spent thou­sands of hours attempt­ing to deter­mine whom, pre­cise­ly, Epstein knew and where, pre­cise­ly, he invest­ed his mon­ey.

Stew­art, who believes that Epstein’s death released him from an oblig­a­tion to con­sid­er the inter­view “on back­ground” and thus anony­mous, revealed that Epstein claimed to have ongo­ing rela­tion­ships with Sau­di Crown Prince Mohammed bin Salman, the dis­graced direc­tor Woody Allen, the for­mer Trump advis­er Steve Ban­non, and the jour­nal­ist Michael Wolff.

Busi­ness Insid­er has also learned that Stew­art’s Times col­league Bowles, who has made a name for her­self skew­er­ing tech oli­garchs and iden­ti­fy­ing Sil­i­con Val­ley cul­tur­al trends, also recent­ly met with Epstein in his Man­hat­tan town­house for an inter­view. It’s unclear if that inter­view was on or off the record, and it’s unclear pre­cise­ly when it occurred. Bowles was list­ed as a con­tribut­ing reporter on a July Times sto­ry fea­tur­ing archi­tec­tur­al and design details about the inte­ri­or of Epstein’s $56 mil­lion town­home, but The Times does not appear to have pub­lished any oth­er report­ing from Bowles’ con­ver­sa­tion with Epstein.

...

The Times and oth­er out­lets have cov­ered Epstein’s efforts, in the wake of his 2008 plea deal, to reha­bil­i­tate his image as a sex offend­er by pay­ing free­lance writ­ers and pub­li­cists to write pos­i­tive sto­ries about him on sites like Huff­Post, Nation­al Review, and Forbes. He also lever­aged a friend­ship with Peg­gy Sie­gal, a pub­li­cist for A‑list celebri­ties, to intro­duce him to a social net­work that includ­ed George Stephanopou­los and Katie Couric.

Town & Coun­try report­ed that Epstein also sought the pub­lic-rela­tions advice of the New York pub­li­cist R. Couri Hay, though Hay nev­er signed him as a client. Hay’s free advice, the mag­a­zine report­ed, was that Epstein should offer him­self up as an exclu­sive inter­view to The Times. Hay declined to com­ment for the record; Sie­gal did not return a mes­sage seek­ing com­ment.

———-

“Jef­frey Epstein was meet­ing with Sil­i­con Val­ley reporters before his arrest, ‘ram­bling’ about all the peo­ple he knew in tech” by John Cook; Busi­ness Insid­er; 08/13/2019

“All three inter­views seem to have touched on Epstein’s rela­tion­ship with Sil­i­con Val­ley. Stew­art wrote that he con­tact­ed Epstein to con­firm a rumor that Epstein was advis­ing Tes­la founder Elon Musk, and both The Infor­ma­tion and Bowles cov­er the tech sec­tor. Stew­art reached out direct­ly to Epstein, but it’s unclear who bro­kered the oth­er meet­ings. The tech focus sug­gests that some­one in Sil­i­con Val­ley may have been try­ing to help Epstein con­nect with reporters.”

Was Hoff­man the mys­tery per­son who may have been bro­ker­ing inter­views with Epstein? Recall that Peter Thiel became an Epstein co-investor in Carbyne911 last year. Might Thiel have been the mys­tery bro­ker? We have no idea, and giv­en the num­ber of con­tacts Epstein has in Sil­i­con Val­ley it’s not like Hoff­man or Thiel are the only sus­pects. As the fol­low­ing arti­cle by Epstein’s biog­ra­ph­er, James B. Stew­art, describes, Epstein was alleged­ly involved with help­ing Elon Musk find a new Tes­la chair­man (some­thing Musk denies). Beyond that, Epstein told Stew­art dur­ing an inter­view last year that he had per­son­al­ly wit­nessed promi­nent tech fig­ures tak­ing drugs and arrang­ing for sex. So when we think about the poten­tial black­mail Epstein’s prob­a­bly had a Sil­i­con Val­ley fig­ures, the num­ber of pos­si­ble fig­ures who may have will­ing­ly or unwill­ing­ly been work­ing to reha­bil­i­tate Epstein’s rep­u­ta­tion is a pret­ty long list:

The New York Times

The Day Jef­frey Epstein Told Me He Had Dirt on Pow­er­ful Peo­ple

By James B. Stew­art
Aug. 12, 2019

Almost exact­ly a year ago, on Aug. 16, 2018, I vis­it­ed Jef­frey Epstein at his cav­ernous Man­hat­tan man­sion.

The over­rid­ing impres­sion I took away from our rough­ly 90-minute con­ver­sa­tion was that Mr. Epstein knew an aston­ish­ing num­ber of rich, famous and pow­er­ful peo­ple, and had pho­tos to prove it. He also claimed to know a great deal about these peo­ple, some of it poten­tial­ly dam­ag­ing or embar­rass­ing, includ­ing details about their sup­posed sex­u­al pro­cliv­i­ties and recre­ation­al drug use.

So one of my first thoughts on hear­ing of Mr. Epstein’s sui­cide was that many promi­nent men and at least a few women must be breath­ing sighs of relief that what­ev­er Mr. Epstein knew, he has tak­en it with him.

Dur­ing our con­ver­sa­tion, Mr. Epstein made no secret of his own scan­dalous past — he’d plead­ed guilty to state charges of solic­it­ing pros­ti­tu­tion from under­age girls and was a reg­is­tered sex offend­er — and acknowl­edged to me that he was a pari­ah in polite soci­ety. At the same time, he seemed unapolo­getic. His very noto­ri­ety, he said, was what made so many peo­ple will­ing to con­fide in him. Every­one, he sug­gest­ed, has secrets and, he added, com­pared with his own, they seemed innocu­ous. Peo­ple con­fid­ed in him with­out feel­ing awk­ward or embar­rassed, he claimed.

I’d nev­er met Mr. Epstein before. I had con­tact­ed him because my col­leagues and I had heard a rumor that he was advis­ing Tesla’s embat­tled chief exec­u­tive, Elon Musk, who was in trou­ble after announc­ing on Twit­ter that he had lined up the fund­ing to take Tes­la pri­vate.

The Secu­ri­ties and Exchange Com­mis­sion began an inves­ti­ga­tion into Mr. Musk’s remarks, which moved mar­kets but didn’t appear to have much basis in fact. There were calls for Mr. Musk to relin­quish his posi­tion as Tesla’s chair­man and for Tes­la to recruit more inde­pen­dent direc­tors. I’d heard that Mr. Epstein was com­pil­ing a list of can­di­dates at Mr. Musk’s behest — and that Mr. Epstein had an email from Mr. Musk autho­riz­ing the search for a new chair­man.

Mr. Musk and Tes­la vehe­ment­ly deny this. “It is incor­rect to say that Epstein ever advised Elon on any­thing,” a spokes­woman for Mr. Musk, Keely Sul­prizio, said Mon­day.

When I con­tact­ed Mr. Epstein, he read­i­ly agreed to an inter­view. The caveat was that the con­ver­sa­tion would be “on back­ground,” which meant I could use the infor­ma­tion as long as I didn’t attribute it direct­ly to him. (I con­sid­er that con­di­tion to have lapsed with his death.)

He insist­ed that I meet him at his house, which I’d seen referred to as the largest sin­gle-fam­i­ly home in Man­hat­tan. This seems plau­si­ble: I ini­tial­ly walked past the build­ing, on East 71st Street, because it looked more like an embassy or muse­um than a pri­vate home. Next to the impos­ing dou­ble doors was a pol­ished brass plaque with the ini­tials “J.E.” and a bell. After I rang, the door was opened by a young woman, her blond hair pulled back in a chignon, who greet­ed me with what sound­ed like an East­ern Euro­pean accent.

I can’t say how old she was, but my guess would be late teens or per­haps 20. Giv­en Mr. Epstein’s past, this struck me as far too close to the line. Why would Mr. Epstein want a reporter’s first impres­sion to be that of a young woman open­ing his door?

The woman led me up a mon­u­men­tal stair­case to a room on the sec­ond floor over­look­ing the Frick muse­um across the street. It was qui­et, the light­ing dim, and the air-con­di­tion­ing was set very low. After a few min­utes, Mr. Epstein bound­ed in, dressed casu­al­ly in jeans and a polo shirt, shook my hand and said he was a big fan of my work. He had a big smile and warm man­ner. He was trim and ener­getic, per­haps from all the yoga he said he was prac­tic­ing. He was unde­ni­ably charis­mat­ic.

Before we left the room he took me to a wall cov­ered with framed pho­tographs. He point­ed to a full-length shot of a man in tra­di­tion­al Arab dress. “That’s M.B.S.,” he said, refer­ring to Mohammed bin Salman, the crown prince of Sau­di Ara­bia. The crown prince had vis­it­ed him many times, and they spoke often, Mr. Epstein said.

He led me to a large room at the rear of the house. There was an expan­sive table with about 20 chairs. Mr. Epstein took a seat at the head, and I sat to his left. He had a com­put­er, a small black­board and a phone to his right. He said he was doing some for­eign-cur­ren­cy trad­ing.

Behind him was a table cov­ered with more pho­tographs. I noticed one of Mr. Epstein with for­mer Pres­i­dent Bill Clin­ton, and anoth­er of him with the direc­tor Woody Allen. Dis­play­ing pho­tos of celebri­ties who had been caught up in sex scan­dals of their own also struck me as odd.

Mr. Epstein avoid­ed specifics about his work for Tes­la. He told me that he had good rea­son to be cryp­tic: Once it became pub­lic that he was advis­ing the com­pa­ny, he’d have to stop doing so, because he was “radioac­tive.” He pre­dict­ed that every­one at Tes­la would deny talk­ing to him or being his friend.

He said this was some­thing he’d become used to, even though it didn’t stop peo­ple from vis­it­ing him, com­ing to his din­ner par­ties or ask­ing him for mon­ey. (That was why, Mr. Epstein told me with­out any trace of irony, he was con­sid­er­ing becom­ing a min­is­ter so that his acquain­tances would be con­fi­dent that their con­ver­sa­tions would be kept con­fi­den­tial.)

If he was ret­i­cent about Tes­la, he was more at ease dis­cussing his inter­est in young women. He said that crim­i­nal­iz­ing sex with teenage girls was a cul­tur­al aber­ra­tion and that at times in his­to­ry it was per­fect­ly accept­able. He point­ed out that homo­sex­u­al­i­ty had long been con­sid­ered a crime and was still pun­ish­able by death in some parts of the world.

Mr. Epstein then mean­dered into a dis­cus­sion of oth­er promi­nent names in tech­nol­o­gy cir­cles. He said peo­ple in Sil­i­con Val­ley had a rep­u­ta­tion for being geeky worka­holics, but that was far from the truth: They were hedo­nis­tic and reg­u­lar users of recre­ation­al drugs. He said he’d wit­nessed promi­nent tech fig­ures tak­ing drugs and arrang­ing for sex (Mr. Epstein stressed that he nev­er drank or used drugs of any kind).

I kept try­ing to steer the con­ver­sa­tion back to Tes­la, but Mr. Epstein remained eva­sive. He said he’d spo­ken to the Saud­is about pos­si­bly invest­ing in Tes­la, but he wouldn’t pro­vide any specifics or names. When I pressed him on the pur­port­ed email from Mr. Musk, he said the email wasn’t from Mr. Musk him­self, but from some­one very close to him. He wouldn’t say who that per­son was. I asked him if that per­son would talk to me, and he said he’d ask. He lat­er said the per­son declined; I doubt he asked.

When I lat­er reflect­ed on our inter­view, I was struck by how lit­tle infor­ma­tion Mr. Epstein had actu­al­ly pro­vid­ed. While I can’t say any­thing he said was an explic­it lie, much of what he said was vague or spec­u­la­tive and couldn’t be proved or dis­proved. He did have at least some ties to Mr. Musk — a wide­ly cir­cu­lat­ed pho­to shows Mr. Musk with Ghis­laine Maxwell, Mr. Epstein’s con­fi­dante and for­mer com­pan­ion, at the 2014 Van­i­ty Fair Oscars par­ty.

“Ghis­laine sim­ply insert­ed her­self behind him in a pho­to he was pos­ing for with­out his knowl­edge,” Ms. Sul­prizio, the spokes­woman for Mr. Musk, said.

It seemed clear Mr. Epstein had embell­ished his role in the Tes­la sit­u­a­tion to enhance his own impor­tance and gain atten­tion — some­thing that now seems to have been a pat­tern.

About a week after that inter­view, Mr. Epstein called and asked if I’d like to have din­ner that Sat­ur­day with him and Woody Allen. I said I’d be out of town. A few weeks after that, he asked me to join him for din­ner with the author Michael Wolff and Don­ald J. Trump’s for­mer advis­er, Steve Ban­non. I declined. (I don’t know if these din­ners actu­al­ly hap­pened. Mr. Ban­non has said he didn’t attend. Mr. Wolff and a spokes­woman for Mr. Allen didn’t respond to requests for com­ment on Mon­day.)

Sev­er­al months passed. Then ear­ly this year Mr. Epstein called to ask if I’d be inter­est­ed in writ­ing his biog­ra­phy. He sound­ed almost plain­tive. I sensed that what he real­ly want­ed was com­pan­ion­ship. As his biog­ra­ph­er, I’d have no choice but to spend hours lis­ten­ing to his saga. Already leery of any fur­ther ties to him, I was relieved I could say that I was already busy with anoth­er book.

...

———-

“The Day Jef­frey Epstein Told Me He Had Dirt on Pow­er­ful Peo­ple” by James B. Stew­art; The New York Times; 08/12/2019

“Mr. Epstein then mean­dered into a dis­cus­sion of oth­er promi­nent names in tech­nol­o­gy cir­cles. He said peo­ple in Sil­i­con Val­ley had a rep­u­ta­tion for being geeky worka­holics, but that was far from the truth: They were hedo­nis­tic and reg­u­lar users of recre­ation­al drugs. He said he’d wit­nessed promi­nent tech fig­ures tak­ing drugs and arrang­ing for sex (Mr. Epstein stressed that he nev­er drank or used drugs of any kind).”

Hav­ing Jef­frey Epstein wit­ness you arrang­ing for sex is prob­a­bly the kind of sit­u­a­tion that will make you high­ly com­pli­ant when it comes to help­ing his rep­u­ta­tion. Or make donations...might that be part of the val­ue Epstein pro­vid­ed for that 2015 din­ner par­ty that was osten­si­bly a fundrais­ing oper­a­tion for Media Lab? Epstein’s pres­ence could pre­sum­ably make any for­mer ‘clients’ of his much more like­ly to open their check­books.

It’s also worth not­ing that Mohammed bin Salman could arguably be con­sid­ered a promi­nent Sil­i­con Val­ley indi­vid­ual giv­en the exten­sive Sau­di invest­ments in Sil­i­con Val­ley com­pa­nies:

...
Before we left the room he took me to a wall cov­ered with framed pho­tographs. He point­ed to a full-length shot of a man in tra­di­tion­al Arab dress. “That’s M.B.S.,” he said, refer­ring to Mohammed bin Salman, the crown prince of Sau­di Ara­bia. The crown prince had vis­it­ed him many times, and they spoke often, Mr. Epstein said.
...

So when Epstein talks about M.B.S. speak­ing to him often and vis­it­ing him many times, while part of the nature of those vis­its could obvi­ous­ly include pros­ti­tu­tion, it’s also very pos­si­ble M.B.S. was using Epstein as a kind of Sil­i­con Val­ley invest­ment front too.

And that’s part of what makes the mys­tery of the iden­ti­ty of Epstein’s main Sil­i­con Val­ley bene­fac­tor so mys­te­ri­ous: there are just way too many viable sus­pects.

Remem­ber when a group of Repub­li­can mem­bers of con­gress stormed into the secure room for high­ly sen­si­tive work (the SCIF) where the House Intel­li­gence Com­mit­tee was hold­ing a impeach­ment hear­ing last month, prompt­ing secu­ri­ty con­cerns over the fact that they brought their cell phones into this room where smart­phones aren’t allowed? Well, here’s an exam­ple of why bring­ing those smart­phones into that room real­ly did pose a very real secu­ri­ty risk. It also hap­pens to be an exam­ple of how smart­phone rep­re­sent a secu­ri­ty risk to pret­ty much any­one:

A new secu­ri­ty flaw was just dis­cov­ered in Google’s wide­ly-used Android oper­at­ing sys­tem for smart­phones. Secu­ri­ty firm Check­marx dis­cov­ered the flaw and cre­at­ed an app demon­strat­ing the large num­ber of ways it can be exploit­ed. It’s like the per­fect flaw for sur­rep­ti­tious tar­get­ed spy­ing or mass spy­ing. The flaw enables any app to poten­tial­ly take con­trol of your smart­phone’s cam­era and micro­phone. Audio and video record­ings and pic­tures can be made and sent back to a com­mand and con­trol serv­er. The attack appears to rely on Google Cam­era app to get around these per­mis­sions. The flaw also allows the attack­er to search through your entire col­lec­tion of pho­tos and videos already stored on the phone and send them back to the serv­er. It can col­lect your GPS loca­tion data too. So it basi­cal­ly turns your smart­phone into the per­fect spy­ing device.

But it gets worse. Because while the use of this flaw would be notice­able if it was being exe­cut­ed while a user was look­ing at their phone (for exam­ple, they would see the video being record­ed in the app), it’s pos­si­ble to use a phone’s prox­im­i­ty sen­sor to deter­mine when the phone is face down when it would be safe to start record­ing with­out the user notic­ing. Anoth­er high­ly oppor­tune time to exploit this vul­ner­a­bil­i­ty is when you are hold­ing your phone up to your ear allow­ing for pic­tures and video to be tak­en of the sur­round­ing room. This is also some­thing apps can detect. Check­marx’s exam­ple mal­ware had both of these capa­bil­i­ties.

Per­haps the worst part of this dis­cov­ered vul­ner­a­bil­i­ty is that it demon­strat­ed how apps were able to eas­i­ly bypass the restric­tions in Android’s oper­at­ing sys­tem that is sup­posed to pre­vent apps from access­ing things like cam­eras or micro­phones with­out users explic­it­ly giv­ing their per­mis­sions. So apps that did­n’t request access to cam­eras and micro­phones could still poten­tial­ly access them on Android phones until this vul­ner­a­bil­i­ty was found. And to upload and videos to the attack­ers’ com­mand and con­trol serv­er only required that the app be giv­en access to phone’s stor­age, which is an extreme­ly com­mon per­mis­sion for apps to request.

At this point we that Android phones built by Google and Sam­sung are vul­ner­a­ble to this attack. We’re also told by Check­marx that Google has pri­vate­ly informed them that oth­er man­u­fac­tur­ers are vul­ner­a­ble, but they haven’t been dis­closed yet. Google issued a state­ment claim­ing that the vul­ner­a­bil­i­ty was addressed on impact­ed Google devices with a July 2019 patch to the Google Cam­era Appli­ca­tion and that patch­es have been made avail­able to all part­ners. Note that in the time­line pro­vid­ed by Check­marx, they informed Google of the vul­ner­a­bil­i­ty on July 4th. So it should have hope­ful­ly been fixed for at least some of the impact­ed peo­ple back in July. At least for Android phones built by Google or Sam­sung. But that still leaves the ques­tion of how long this kind of vul­ner­a­bil­i­ty has been exploitable:

Ars Tech­ni­ca

Google & Sam­sung fix Android spy­ing flaw. Oth­er mak­ers may still be vul­ner­a­ble
Cam­era and mic could be con­trolled by any app, no per­mis­sion required.

Dan Good­in — 11/19/2019, 6:32 AM

Until recent­ly, weak­ness­es in Android cam­era apps from Google and Sam­sung made it pos­si­ble for rogue apps to record video and audio and take images and then upload them to an attack­er-con­trolled server—without any per­mis­sions to do so. Cam­era apps from oth­er man­u­fac­tur­ers may still be sus­cep­ti­ble.

The weak­ness, which was dis­cov­ered by researchers from secu­ri­ty firm Check­marx, rep­re­sent­ed a poten­tial pri­va­cy risk to high-val­ue tar­gets, such as those preyed upon by nation-spon­sored spies. Google care­ful­ly designed its Android oper­at­ing sys­tem to bar apps from access­ing cam­eras and micro­phones with­out explic­it per­mis­sion from end users. An inves­ti­ga­tion pub­lished Tues­day showed it was triv­ial to bypass those restric­tions. The inves­ti­ga­tion found that an app need­ed no per­mis­sions at all to cause the cam­era to shoot pic­tures and record video and audio. To upload the images and video—or any oth­er image and video stored on the phone—to an attack­er-con­trolled serv­er, an app need­ed only per­mis­sion to access stor­age, which is among one of the most com­mon­ly giv­en usage rights.

The weak­ness, which is tracked as CVE-2019–2234, also allowed would-be attack­ers to track the phys­i­cal loca­tion of the device, assum­ing GPS data was embed­ded into images or videos. Google closed the eaves­drop­ping hole in its Pix­el line of devices with a cam­era update that became avail­able in July. Check­marx said Sam­sung has also fixed the vul­ner­a­bil­i­ty, although it was­n’t clear when that hap­pened. Check­marx said Google has indi­cat­ed that Android phones from oth­er man­u­fac­tur­ers may also be vul­ner­a­ble. The spe­cif­ic mak­ers and mod­els haven’t been dis­closed.

“The abil­i­ty for an appli­ca­tion to retrieve input from the cam­era, micro­phone, and GPS loca­tion is con­sid­ered high­ly inva­sive by Google them­selves,” Check­marx Direc­tor of Secu­ri­ty Research Erez Yalon wrote in Tues­day’s analy­sis. “As a result, AOSP cre­at­ed a spe­cif­ic set of per­mis­sions that an appli­ca­tion must request from the user.”

To demon­strate the risk, Check­marx devel­oped a proof-of-con­cept rogue app that exploit­ed the weak­ness. It mas­quer­ad­ed as a sim­ple weath­er app. Hid­den inside were func­tions that could:

* Take pic­tures and record videos, even when the phone was locked, the screen was off, or the app was closed
* Pull GPS data embed­ded into any pho­to or video stored on the phone
* Eaves­drop and record two-way phone con­ver­sa­tions and simul­ta­ne­ous­ly record video or take images
* Silence the cam­era shut­ter to make the spy­ing hard­er to detect
* Trans­fer any pho­to or video stored on the phone to an attack­er-con­trolled serv­er
* List and down­load any JPG image or MP4 video stored on the phone’s SD card

An attack would­n’t be com­plete­ly sur­rep­ti­tious. The screen of an exploit­ed device would dis­play the cam­era as it record­ed video or shot an image. That would tip off any­one who was look­ing at the hand­set at the time the attack was being car­ried out. Still, the attack would be able to cap­ture video, sound, and images at times when a phone dis­play was out of eye­sight, such as when the device was placed screen down. The app was able to use the prox­im­i­ty sen­sor to deter­mine when the device is face down.

Check­marx’s PoC app was also able to use a phone’s prox­im­i­ty sen­sor to detect when it was held to a tar­get’s ear, as often hap­pens dur­ing phone calls. The app was able to record both sides of the con­ver­sa­tion. It could also record video or take images, a use­ful capa­bil­i­ty in the event the back of the phone was fac­ing a white­board or some­thing else of inter­est to an attack­er. Check­marx’s report includes a video demon­strat­ing the capa­bil­i­ties of the PoC app.

In a state­ment, Google offi­cials wrote: “We appre­ci­ate Check­marx bring­ing this to our atten­tion and work­ing with Google and Android part­ners to coor­di­nate dis­clo­sure. The issue was addressed on impact­ed Google devices via a Play Store update to the Google Cam­era Appli­ca­tion in July 2019. A patch has also been made avail­able to all part­ners.”

Sam­sung offi­cials wrote: “Since being noti­fied of this issue by Google, we have sub­se­quent­ly released patch­es to address all Sam­sung device mod­els that may be affect­ed. We val­ue our part­ner­ship with the Android team that allowed us to iden­ti­fy and address this mat­ter direct­ly.”

...

Check­marx said Google has pri­vate­ly indi­cat­ed that oth­er mak­ers of Android phones besides Sam­sung may also be vul­ner­a­ble. Google’s state­ment did­n’t direct­ly con­firm this or say if any oth­er man­u­fac­tur­ers have installed an update.

In an email, Check­marx’s Yalon said it was­n’t clear why apps could access the cam­era with­out the user pro­vid­ing per­mis­sion. He spec­u­lat­ed that the weak­ness may be the result of Google mak­ing the cam­era work with the voice-acti­vat­ed Google Assis­tant and oth­er man­u­fac­tur­ers fol­low­ing suit.

Users of Pix­el phones can con­firm they aren’t vul­ner­a­ble by access­ing Apps and Noti­fi­ca­tions from the set­tings menu, choos­ing Cam­era > Advanced > and App details. The screen should show that the app has been updat­ed since July (and ide­al­ly much more recent­ly than that).

Check­ing if oth­er Android phones are sus­cep­ti­ble will be dif­fi­cult for most users. Those who are more tech­ni­cal­ly skilled can run the fol­low­ing com­mand:

$ adb shell am start-activ­i­ty ‑n
com.google.android.GoogleCamera/com.android.camera.CameraActivity –ez
extra_turn_screen_on true ‑a android.media.action.VIDEO_CAMERA –ez
android.intent.extra.USE_FRONT_CAMERA true

The above com­mand will force the phone to take video. The fol­low­ing com­mand will force the phone to take a pho­to:

$ adb shell am start-activ­i­ty ‑n
com.google.android.GoogleCamera/com.android.camera.CameraActivity –ez
extra_turn_screen_on true ‑a android.media.action.STILL_IMAGE_CAMERA -
‑ez android.intent.extra.USE_FRONT_CAMERA true –ei
android.intent.extra.TIMER_DURATION_SECONDS 3

The skill and luck required to make the attack work reli­ably and with­out detec­tion are high enough that this type of exploit isn’t like­ly to be used against the vast major­i­ty of Android users. Still, the ease of sneak­ing mali­cious apps into the Google Play store sug­gests it would­n’t be hard for a deter­mined and sophis­ti­cat­ed attack­er to pull off some­thing like this. No won­der phones and oth­er elec­tron­ics are barred from SCIFs and oth­er sen­si­tive envi­ron­ments.

———-

“Google & Sam­sung fix Android spy­ing flaw. Oth­er mak­ers may still be vul­ner­a­ble” by Dan Good­in; Ars Tech­ni­ca; 11/19/2019

“The skill and luck required to make the attack work reli­ably and with­out detec­tion are high enough that this type of exploit isn’t like­ly to be used against the vast major­i­ty of Android users. Still, the ease of sneak­ing mali­cious apps into the Google Play store sug­gests it would­n’t be hard for a deter­mined and sophis­ti­cat­ed attack­er to pull off some­thing like this. No won­der phones and oth­er elec­tron­ics are barred from SCIFs and oth­er sen­si­tive envi­ron­ments.”

Have sophis­ti­cat­ed attack­ers been using this vul­ner­a­bil­i­ty all along? We don’t know, but it did­n’t sound like Check­marx had a very hard time dis­cov­er­ing this. And giv­en how Check­marx was able to build their proof-of-con­cept app to only oper­ate when the phone was either face down or being held up to some­one’s ear, it’s pos­si­ble this has been a wide­ly used hack that no one noticed:

...
To demon­strate the risk, Check­marx devel­oped a proof-of-con­cept rogue app that exploit­ed the weak­ness. It mas­quer­ad­ed as a sim­ple weath­er app. Hid­den inside were func­tions that could:

* Take pic­tures and record videos, even when the phone was locked, the screen was off, or the app was closed
* Pull GPS data embed­ded into any pho­to or video stored on the phone
* Eaves­drop and record two-way phone con­ver­sa­tions and simul­ta­ne­ous­ly record video or take images
* Silence the cam­era shut­ter to make the spy­ing hard­er to detect
* Trans­fer any pho­to or video stored on the phone to an attack­er-con­trolled serv­er
* List and down­load any JPG image or MP4 video stored on the phone’s SD card

An attack would­n’t be com­plete­ly sur­rep­ti­tious. The screen of an exploit­ed device would dis­play the cam­era as it record­ed video or shot an image. That would tip off any­one who was look­ing at the hand­set at the time the attack was being car­ried out. Still, the attack would be able to cap­ture video, sound, and images at times when a phone dis­play was out of eye­sight, such as when the device was placed screen down. The app was able to use the prox­im­i­ty sen­sor to deter­mine when the device is face down.

Check­marx’s PoC app was also able to use a phone’s prox­im­i­ty sen­sor to detect when it was held to a tar­get’s ear, as often hap­pens dur­ing phone calls. The app was able to record both sides of the con­ver­sa­tion. It could also record video or take images, a use­ful capa­bil­i­ty in the event the back of the phone was fac­ing a white­board or some­thing else of inter­est to an attack­er. Check­marx’s report includes a video demon­strat­ing the capa­bil­i­ties of the PoC app.
...

So if you have an Android phone with some ques­tion­able apps , espe­cial­ly phones not man­u­fac­tured by Google or Sam­sung and there­fore poten­tial­ly still vul­ner­a­ble, it might be worth run­ning that app and then lay­ing the phone down a glass sur­face so you can still see what’s hap­pen­ing on the phone’s screen.

Also note how Check­marx’s report isn’t just dis­clos­ing this vul­ner­a­bil­i­ty exploit­ed via the Google Cam­era app. It’s also a reminder that when apps are access to a phone’s stor­age device, there’s noth­ing real­ly stop­ping those apps from root­ing through all of the oth­er data on your phone’s stor­age card. Like all your pho­tos and videos. And then upload­ing them to a serv­er:

...
The weak­ness, which was dis­cov­ered by researchers from secu­ri­ty firm Check­marx, rep­re­sent­ed a poten­tial pri­va­cy risk to high-val­ue tar­gets, such as those preyed upon by nation-spon­sored spies. Google care­ful­ly designed its Android oper­at­ing sys­tem to bar apps from access­ing cam­eras and micro­phones with­out explic­it per­mis­sion from end users. An inves­ti­ga­tion pub­lished Tues­day showed it was triv­ial to bypass those restric­tions. The inves­ti­ga­tion found that an app need­ed no per­mis­sions at all to cause the cam­era to shoot pic­tures and record video and audio. To upload the images and video—or any oth­er image and video stored on the phone—to an attack­er-con­trolled serv­er, an app need­ed only per­mis­sion to access stor­age, which is among one of the most com­mon­ly giv­en usage rights.
...

As Check­marx describes in their report, when you give an app in Android access to the stor­age on the device, you aren’t just giv­ing it access to its own stored data. You are giv­ing the app access to every­thing stored on that SD card:

Check­marx

How Attack­ers Could Hijack Your Android Cam­era to Spy on You

Nov 19, 2019 by Erez Yalon

In today’s dig­i­tal­ly-con­nect­ed soci­ety, smart­phones have become an exten­sion of us. Advanced cam­era and video capa­bil­i­ties in par­tic­u­lar are play­ing a mas­sive role in this, as users are able to quick­ly take out their phones and cap­ture any moment in real-time with the sim­ple click of a but­ton. How­ev­er, this presents a dou­ble-edged sword as these mobile devices are con­stant­ly col­lect­ing, stor­ing, and shar­ing var­i­ous types of data – with and with­out our know­ing – mak­ing our devices gold­mines for attack­ers.

In order to bet­ter under­stand how smart­phone cam­eras may be open­ing users up to pri­va­cy risks, the Check­marx Secu­ri­ty Research Team cracked into the appli­ca­tions them­selves that con­trol these cam­eras to iden­ti­fy poten­tial abuse sce­nar­ios. Hav­ing a Google Pix­el 2 XL and Pix­el 3 on-hand, our team began research­ing the Google Cam­era app [1], ulti­mate­ly find­ing mul­ti­ple con­cern­ing vul­ner­a­bil­i­ties stem­ming from per­mis­sion bypass issues. After fur­ther dig­ging, we also found that these same vul­ner­a­bil­i­ties impact the cam­era apps of oth­er smart­phone ven­dors in the Android ecosys­tem – name­ly Sam­sung – pre­sent­ing sig­nif­i­cant impli­ca­tions to hun­dreds-of-mil­lions of smart­phone users.

In this blog, we’ll explain the vul­ner­a­bil­i­ties dis­cov­ered (CVE-2019–2234), pro­vide details of how they were exploit­ed, explain the con­se­quences, and note how users can safe­guard their devices. This blog is also accom­pa­nied by a proof-of-con­cept (PoC) video, as well as a tech­ni­cal report of the find­ings that were shared with Google, Sam­sung, and oth­er Android-based smart­phone OEMs.

Google & Sam­sung Cam­era Vul­ner­a­bil­i­ties

After a detailed analy­sis of the Google Cam­era app, our team found that by manip­u­lat­ing spe­cif­ic actions and intents [2], an attack­er can con­trol the app to take pho­tos and/or record videos through a rogue appli­ca­tion that has no per­mis­sions to do so. Addi­tion­al­ly, we found that cer­tain attack sce­nar­ios enable mali­cious actors to cir­cum­vent var­i­ous stor­age per­mis­sion poli­cies, giv­ing them access to stored videos and pho­tos, as well as GPS meta­da­ta embed­ded in pho­tos, to locate the user by tak­ing a pho­to or video and pars­ing the prop­er EXIF data [3]. This same tech­nique also applied to Samsung’s Cam­era app.

In doing so, our researchers deter­mined a way to enable a rogue appli­ca­tion to force the cam­era apps to take pho­tos and record video, even if the phone is locked or the screen is turned off. Our researchers could do the same even when a user was is in the mid­dle of a voice call.

The Impli­ca­tions

The abil­i­ty for an appli­ca­tion to retrieve input from the cam­era, micro­phone, and GPS loca­tion is con­sid­ered high­ly inva­sive by Google them­selves. As a result, AOSP cre­at­ed a spe­cif­ic set of per­mis­sions that an appli­ca­tion must request from the user. Since this was the case, Check­marx researchers designed an attack sce­nario that cir­cum­vents this per­mis­sion pol­i­cy by abus­ing the Google Cam­era app itself, forc­ing it to do the work on behalf of the attack­er.

It is known that Android cam­era appli­ca­tions usu­al­ly store their pho­tos and videos on the SD card. Since pho­tos and videos are sen­si­tive user infor­ma­tion, in order for an appli­ca­tion to access them, it needs spe­cial per­mis­sions: stor­age per­mis­sions. Unfor­tu­nate­ly, stor­age per­mis­sions are very broad and these per­mis­sions give access to the entire SD card. There are a large num­ber of appli­ca­tions, with legit­i­mate use-cas­es, that request access to this stor­age, yet have no spe­cial inter­est in pho­tos or videos. In fact, it’s one of the most com­mon request­ed per­mis­sions observed.

This means that a rogue appli­ca­tion can take pho­tos and/or videos with­out spe­cif­ic cam­era per­mis­sions, and it only needs stor­age per­mis­sions to take things a step fur­ther and fetch pho­tos and videos after being tak­en. Addi­tion­al­ly, if the loca­tion is enabled in the cam­era app, the rogue appli­ca­tion also has a way to access the cur­rent GPS posi­tion of the phone and user.

Of course, a video also con­tains sound. It was inter­est­ing to prove that a video could be ini­ti­at­ed dur­ing a voice call. We could eas­i­ly record the receiver’s voice dur­ing the call and we could record the caller’s voice as well.

...
———-

“How Attack­ers Could Hijack Your Android Cam­era to Spy on You” by Erez Yalon; Check­marx; 11/19/2019

“It is known that Android cam­era appli­ca­tions usu­al­ly store their pho­tos and videos on the SD card. Since pho­tos and videos are sen­si­tive user infor­ma­tion, in order for an appli­ca­tion to access them, it needs spe­cial per­mis­sions: stor­age per­mis­sions. Unfor­tu­nate­ly, stor­age per­mis­sions are very broad and these per­mis­sions give access to the entire SD card. There are a large num­ber of appli­ca­tions, with legit­i­mate use-cas­es, that request access to this stor­age, yet have no spe­cial inter­est in pho­tos or videos. In fact, it’s one of the most com­mon request­ed per­mis­sions observed.”

So while this recent­ly dis­closed vul­ner­a­bil­i­ty is pri­mar­i­ly focused on how the Google Cam­era app had this mas­sive vul­ner­a­bil­i­ty that allowed for the hijack­ing of cam­eras and micro­phones, it’s also a remind that all of the con­tents of your Smart­phone’s SD cards are poten­tial­ly avail­able to any app on your phone as long as those apps have been giv­en the “Stor­age” per­mis­sions. And that’s not just a vul­ner­a­bil­i­ty that needs to be fixed. It’s a basic part of how the Android oper­at­ing sys­tem works.

Also don’t for­get that Google was start­ed with seed fund­ing from the CIA. So when we learn about these kinds of vul­ner­a­bil­i­ties that are almost tai­lor made for spies, maybe that’s what they are.

It’s all a reminder that mod­ern tech­nol­o­gy regime is pred­i­cat­ed on sys­tems of trust. Trust in soft­ware and hard­ware devel­op­ers that the vast major­i­ty of users can’t real­is­ti­cal­ly have a basis for giv­ing and yet must giv­en in order to use the tech­nol­o­gy. In oth­er words, our mod­ern tech­nol­o­gy regime is pred­i­cat­ed on sys­tems of untrust­wor­thy trust. Which seems like a pret­ty huge secu­ri­ty vul­ner­a­bil­i­ty.

Here’s a sto­ry about Cam­bridge Ana­lyt­i­ca that’s real­ly about a much larg­er sto­ry about Cam­bridge Ana­lyt­i­ca that’s going to be unfold­ing over the com­ing months: a large leak over over 100,000 Cam­bridge Ana­lyt­i­ca doc­u­ments has start­ed trick­ling online from the anony­mous @HindsightFiles twit­ter account. The files came from the emails accounts and hard dri­ves of Brit­tany Kaiser. Recall how Kaiser, the direc­tor of busi­ness devel­op­ment at SCL between Feb­ru­ary 2015 and Jan­u­ary of 2018, has already come for­ward and claimed that the ~87 mil­lion esti­mate of the num­ber of peo­ple who had their Face­book pro­file infor­ma­tion col­lect­ed by Cam­bridge Ana­lyt­i­ca is too low and the real num­ber is “much greater”. We don’t know yet if Kaiser is the direct source of these anony­mous leaks, but it’s her files get­ting leaked. Kaiser has decid­ed to speak out pub­licly about the full scope of Cam­bridge Ana­lyt­i­ca’s activ­i­ties fol­low­ing the elec­tion in the UK last month. The way she puts it, her cache of files con­tains thou­sands and thou­sands more pages which showed a “breadth and depth of the work” that went “way beyond what peo­ple think they know about ‘the Cam­bridge Ana­lyt­i­ca scan­dal’”. The files also turn out to be the same files sub­poe­naed by the Mueller inves­ti­ga­tion.

So what new infor­ma­tion has been released so far? Well, it’s quite a tease: we’re told the doc­u­ments are going to relate to Cam­bridge Ana­lyt­i­ca’s work in 68 coun­tries. And the “indus­tri­al scale” nature of the oper­a­tion is going to be laid bare. The doc­u­ment release began on New Year’s Day and includ­ed mate­ri­als on elec­tions in Malaysia, Kenya, and Brazil. The files also include mate­r­i­al that sug­gests Cam­bridge Ana­lyt­i­ca was work­ing for a polit­i­cal par­ty in Ukraine in 2017. We don’t yet know which par­ty.

Unsur­pris­ing­ly, there’s also a Dark Mon­ey angle to the sto­ry. The doc­u­ments include emails between major Trump donors dis­cussing ways of obscur­ing the source of their dona­tions through a series of dif­fer­ent finan­cial vehi­cles. So the unlim­it­ed secret financ­ing of polit­i­cal cam­paigns allowed by US elec­tion law includes the secret financ­ing of secret sophis­ti­cat­ed social media psy­cho­log­i­cal manip­u­la­tion cam­paigns too. Sur­prise. Only some of the 100,000+ doc­u­ments have been leaked so far and more are set to be released in com­ing months. So the @HindsightFiles twit­ter account is going to be one to watch:

The Guardian

Fresh Cam­bridge Ana­lyt­i­ca leak ‘shows glob­al manip­u­la­tion is out of con­trol’

Company’s work in 68 coun­tries laid bare with release of more than 100,000 doc­u­ments

Car­ole Cad­wal­ladr
Sat 4 Jan 2020 11.55 EST
Last mod­i­fied on Mon 6 Jan 2020 06.37 EST

An explo­sive leak of tens of thou­sands of doc­u­ments from the defunct data firm Cam­bridge Ana­lyt­i­ca is set to expose the inner work­ings of the com­pa­ny that col­lapsed after the Observ­er revealed it had mis­ap­pro­pri­at­ed 87 mil­lion Face­book pro­files.

More than 100,000 doc­u­ments relat­ing to work in 68 coun­tries that will lay bare the glob­al infra­struc­ture of an oper­a­tion used to manip­u­late vot­ers on “an indus­tri­al scale” are set to be released over the next months.

It comes as Christo­pher Steele, the ex-head of MI6’s Rus­sia desk and the intel­li­gence expert behind the so-called “Steele dossier” into Trump’s rela­tion­ship with Rus­sia, said that while the com­pa­ny had closed down, the fail­ure to prop­er­ly pun­ish bad actors meant that the prospects for manip­u­la­tion of the US elec­tion this year were even worse.

The release of doc­u­ments began on New Year’s Day on an anony­mous Twit­ter account, @HindsightFiles, with links to mate­r­i­al on elec­tions in Malaysia, Kenya and Brazil. The doc­u­ments were revealed to have come from Brit­tany Kaiser, an ex-Cam­bridge Ana­lyt­i­ca employ­ee turned whistle­blow­er, and to be the same ones sub­poe­naed by Robert Mueller’s inves­ti­ga­tion into Russ­ian inter­fer­ence in the 2016 pres­i­den­tial elec­tion.

Kaiser, who starred in the Oscar-short­list­ed Net­flix doc­u­men­tary The Great Hack, decid­ed to go pub­lic after last month’s elec­tion in Britain. “It’s so abun­dant­ly clear our elec­toral sys­tems are wide open to abuse,” she said. “I’m very fear­ful about what is going to hap­pen in the US elec­tion lat­er this year, and I think one of the few ways of pro­tect­ing our­selves is to get as much infor­ma­tion out there as pos­si­ble.”

The doc­u­ments were retrieved from her email accounts and hard dri­ves, and though she hand­ed over some mate­r­i­al to par­lia­ment in April 2018, she said there were thou­sands and thou­sands more pages which showed a “breadth and depth of the work” that went “way beyond what peo­ple think they know about ‘the Cam­bridge Ana­lyt­i­ca scan­dal’”.

Steele made a rare pub­lic inter­ven­tion to com­ment on the leaks. He said that while he didn’t know what was in them, the con­text couldn’t be more impor­tant because “on our cur­rent tra­jec­to­ry these prob­lems are like­ly to get worse, not bet­ter, and with cru­cial 2020 elec­tions in Amer­i­ca and else­where approach­ing, this is a very scary prospect. Some­thing rad­i­cal needs to be done about it, and fast.”

...

Kaiser said the Face­book data scan­dal was part of a much big­ger glob­al oper­a­tion that worked with gov­ern­ments, intel­li­gence agen­cies, com­mer­cial com­pa­nies and polit­i­cal cam­paigns to manip­u­late and influ­ence peo­ple, and that raised huge nation­al secu­ri­ty impli­ca­tions.

The unpub­lished doc­u­ments con­tain mate­r­i­al that sug­gests the firm was work­ing for a polit­i­cal par­ty in Ukraine in 2017 even while under inves­ti­ga­tion as part of Mueller’s inquiry and emails that Kaiser says describe how the firm helped devel­op a “sophis­ti­cat­ed infra­struc­ture of shell com­pa­nies that were designed to fun­nel dark mon­ey into pol­i­tics”.

“There are emails between these major Trump donors dis­cussing ways of obscur­ing the source of their dona­tions through a series of dif­fer­ent finan­cial vehi­cles. These doc­u­ments expose the entire dark mon­ey machin­ery behind US pol­i­tics.” The same machin­ery, she says, was deployed in oth­er coun­tries that Cam­bridge Ana­lyt­i­ca worked in, includ­ing, she claims, Britain.

Emma Bri­ant, an aca­d­e­m­ic at Bard Col­lege, New York, who spe­cialis­es in inves­ti­gat­ing pro­pa­gan­da and has had access to some of the doc­u­ments for research, said that what had been revealed was “the tip of the ice­berg”.

“The doc­u­ments reveal a much clear­er idea of what actu­al­ly hap­pened in the 2016 US pres­i­den­tial elec­tion, which has a huge bear­ing on what will hap­pen in 2020. It’s the same peo­ple involved who we know are build­ing on these same tech­niques,” she said.

“There’s evi­dence of real­ly quite dis­turb­ing exper­i­ments on Amer­i­can vot­ers, manip­u­lat­ing them with fear-based mes­sag­ing, tar­get­ing the most vul­ner­a­ble, that seems to be con­tin­u­ing. This is an entire glob­al indus­try that’s out of con­trol but what this does is lay out what was hap­pen­ing with this one com­pa­ny.”

———-

“Fresh Cam­bridge Ana­lyt­i­ca leak ‘shows glob­al manip­u­la­tion is out of con­trol’” by Car­ole Cad­wal­ladr; The Guardian; 01/04/2020

“The release of doc­u­ments began on New Year’s Day on an anony­mous Twit­ter account, @HindsightFiles, with links to mate­r­i­al on elec­tions in Malaysia, Kenya and Brazil. The doc­u­ments were revealed to have come from Brit­tany Kaiser, an ex-Cam­bridge Ana­lyt­i­ca employ­ee turned whistle­blow­er, and to be the same ones sub­poe­naed by Robert Mueller’s inves­ti­ga­tion into Russ­ian inter­fer­ence in the 2016 pres­i­den­tial elec­tion.”

So the trove of Kaiser’s doc­u­ments hand­ed over to the Mueller team are set to be released in com­ing months. That’s excit­ing. Espe­cial­ly since she’s describ­ing the full scope of the Cam­bridge Ana­lyt­i­ca oper­a­tion as includ­ing the coor­di­na­tion of gov­ern­ments and intel­li­gence agen­cies, in addi­tion to the polit­i­cal cam­paigns we already knew about. Hope­ful­ly we get to learn about about which Ukrain­ian polit­i­cal par­ty Cam­bridge Ana­lyt­i­ca was work­ing with in 2017:

...
The doc­u­ments were retrieved from her email accounts and hard dri­ves, and though she hand­ed over some mate­r­i­al to par­lia­ment in April 2018, she said there were thou­sands and thou­sands more pages which showed a “breadth and depth of the work” that went “way beyond what peo­ple think they know about ‘the Cam­bridge Ana­lyt­i­ca scan­dal’”.

...

Kaiser said the Face­book data scan­dal was part of a much big­ger glob­al oper­a­tionthat worked with gov­ern­ments, intel­li­gence agen­cies, com­mer­cial com­pa­nies and polit­i­cal cam­paigns to manip­u­late and influ­ence peo­ple, and that raised huge nation­al secu­ri­ty impli­ca­tions.

The unpub­lished doc­u­ments con­tain mate­r­i­al that sug­gests the firm was work­ing for a polit­i­cal par­ty in Ukraine in 2017 even while under inves­ti­ga­tion as part of Mueller’s inquiry and emails that Kaiser says describe how the firm helped devel­op a “sophis­ti­cat­ed infra­struc­ture of shell com­pa­nies that were designed to fun­nel dark mon­ey into pol­i­tics”.

“There are emails between these major Trump donors dis­cussing ways of obscur­ing the source of their dona­tions through a series of dif­fer­ent finan­cial vehi­cles. These doc­u­ments expose the entire dark mon­ey machin­ery behind US pol­i­tics.” The same machin­ery, she says, was deployed in oth­er coun­tries that Cam­bridge Ana­lyt­i­ca worked in, includ­ing, she claims, Britain.
...

So with much greater scope of the Cam­bridge Ana­lyt­i­ca oper­a­tion in mind, here’s a Gray­zone piece from 2018 that describes “Project Tita­nia”, the name for an oper­a­tion focused on psy­cho­log­i­cal­ly pro­fil­ing the Yemeni pop­u­la­tion for the US mil­i­tary. The arti­cle is based on doc­u­ments that describe SCL’s work as a mil­i­tary con­trac­tor in coun­tries around the world and includes some ear­li­er work SCL did in Ukraine. The work is so ear­ly it either pre­ced­ed the for­mal incor­po­ra­tion of SCL or must have been one of SCL’s very first projects. Because accord­ing to the inter­nal SCL doc­u­ments they obtained, SCL was work­ing on the pro­mot­ing the “Orange Rev­o­lu­tion” in Ukraine back in late 2004. SCL was start­ed in 2005. So Ukraine appears to have been one of SCL’s very first projects. The doc­u­ments obtained by the Gray­zone Project also describe oper­a­tions across the Mid­dle East as a US and UK counter-insur­gency con­trac­tor, includ­ing an oper­a­tion in Iran in 2009. It points towards a key con­text to keep in mind as Kaiser’s 100,000+ doc­u­ments are released in com­ing months: while much of what Cam­bridge Ana­lyt­i­ca and its SCL par­ent com­pa­ny were doing in those 68 coun­tries was prob­a­bly done at the behest of pri­vate clients, we can’t for­get that SCL has a long his­to­ry as a mil­i­tary con­trac­tor too. The US and UK mil­i­tary and intel­li­gence agen­cies were prob­a­bly clients in most of those cas­es, but it’s also prob­a­bly not lim­it­ed to the US and UK. As Kaiser warns us, this is glob­al oper­a­tion. And these ser­vices have been up for sale since as far back as Ukraine’s Orange Rev­o­lu­tion:

The Gray­zone Project

Exclu­sive Leaked Docs Expose Yemen-Based Counter-Insur­gency Pro­gram by Cam­bridge Ana­lyt­i­ca Par­ent Com­pa­ny SCL

Part one of a two part inves­ti­ga­tion into Cam­bridge Analytica/SCL’s glob­al oper­a­tions

By Max Blu­men­thal
May 23, 2018

Inter­nal doc­u­ments exclu­sive­ly obtained by the Gray­zone Project and embed­ded at the end of this arti­cle show how Cam­bridge Analytica’s UK-based par­ent com­pa­ny, SCL group, con­duct­ed a sur­veil­lance oper­a­tion in Yemen, using psy­cho­log­i­cal pro­fil­ing, “strate­gic com­mu­ni­ca­tions cam­paigns,” and infil­tra­tion of for­eign oper­a­tives into indige­nous com­mu­ni­ties through unwit­ting local part­ners whom they were instruct­ed to deceive.

The SCL doc­u­ments describe “a research and analy­sis study under­tak­en by Strate­gic Com­mu­ni­ca­tion Lab­o­ra­to­ries (SCL) on behalf of Archimedes,” a US-based mil­i­tary con­trac­tor. The name of the oper­a­tion was “Project Tita­nia.” It relied heav­i­ly on decep­tion to gain access to the local pop­u­la­tion, order­ing project oper­a­tives to devel­op a “cov­er sto­ry” that placed their pres­ence in the coun­try in a more inno­cent light.

The geo­graph­ic tar­gets of the project were Yemen’s Hadra­mout and Marib provinces. These regions have served as orga­ni­za­tion­al bases for Al Qae­da in the Ara­bi­an Penin­su­la, and were at the time in the crosshairs of then-Pres­i­dent Barack Obama’s drone assas­si­na­tion pro­gram.

Many of the meth­ods of sur­veil­lance and manip­u­la­tion revealed in these SCL doc­u­ments close­ly mir­ror the tac­tics that were lat­er applied in West­ern elec­toral con­tests. And when these tac­tics were exposed in ear­ly 2018, they ignit­ed a polit­i­cal firestorm.

The data and behav­ioral ana­lyt­ics firm Cam­bridge Ana­lyt­i­ca closed shop this May fol­low­ing dam­ag­ing rev­e­la­tions that it obtained the raw data of over 80 mil­lion Face­book users dur­ing the 2016 pres­i­den­tial elec­tion cam­paign, and exploit­ed this infor­ma­tion to influ­ence the out­come of numer­ous polit­i­cal cam­paigns. The dam­age spread across the Atlantic, to Cam­bridge Analytica’s UK-based par­ent com­pa­ny, SCL group, forc­ing it to shut­ter its oper­a­tions as well.

Cam­bridge Ana­lyt­i­ca was par­tial­ly owned by Steve Ban­non, the for­mer man­ag­er of Trump’s 2016 pres­i­den­tial cam­paign and ex-White House chief of staff. It was co-owned by Bannon’s main finan­cial angel at the time, reclu­sive right-wing tech bil­lion­aire Robert Mer­cer, and his daugh­ter, Rebekah Mer­cer, who served as its vice pres­i­dent.

Cam­bridge Ana­lyt­i­ca debuted its work with a series of Repub­li­can Get Out the Vote efforts in 2014, deploy­ing the psy­cho­log­i­cal pro­fil­ing tech­niques that have become the company’s bread and but­ter. “Its dirty lit­tle secret was that there was no one Amer­i­can involved in [the 2016 effort], that it was a de fac­to for­eign agent, work­ing on an Amer­i­can elec­tion,” for­mer Cam­bridge Ana­lyt­i­ca research direc­tor Christo­pher Wylie revealed.

Dur­ing the 2016 pres­i­den­tial cam­paign, Cam­bridge Ana­lyt­i­ca played a cen­tral role in Trump’s dig­i­tal out­reach efforts. In a hid­den cam­era inves­ti­ga­tion by the UK’s Chan­nel 4, Cam­bridge Ana­lyt­i­ca exec­u­tives took cred­it for gen­er­at­ing sev­er­al endur­ing lines of attack against Hillary Clin­ton. “We just put infor­ma­tion into the blood­stream of the inter­net and then watch it grow, give it a lit­tle push every now and again over time to watch it take shape,” one exec­u­tive boast­ed. “And so this stuff infil­trates the online com­mu­ni­ty, but with no brand­ing, so it’s unat­trib­ut­able, untrack­able.”

SCL Group and ‘Project Tita­nia‘

The eth­i­cal­ly dubi­ous tac­tics that Cam­bridge Ana­lyt­i­ca rolled out dur­ing the 2016 cam­paign had been honed by its par­ent com­pa­ny, the Lon­don-based SCL Group, in an array of influ­ence oper­a­tions in con­flict zones and Third World elec­tion con­tests.

Found­ed in 2005, SCL spe­cial­izes in what com­pa­ny lit­er­a­ture has described as “influ­ence oper­a­tions” and “psy­cho­log­i­cal war­fare” around the globe. An SCL brochure leaked to the BBC revealed how the firm exac­er­bat­ed eth­nic ten­sions in Latvia to assist their client in 2006.

A year lat­er, the firm orches­trat­ed “anti-elec­tion” ral­lies to sup­press the oppo­si­tion vote in Nigeria’s 2007 pres­i­den­tial elec­tion. Then, in 2010, accord­ing to the BBC, SCL ini­ti­at­ed an “ambi­tious cam­paign of polit­i­cal graf­fi­ti” that “osten­si­bly came from the youth,” enabling its client to “claim cred­it for lis­ten­ing to a ‘unit­ed youth.’”

SCL has also applied its influ­ence in Ukraine, first as part of the broad­er pub­lic rela­tions cam­paign dur­ing the country’s NATO-backed 2004 “Orange Rev­o­lu­tion,” then in Ukraine’s con­test­ed Don­bas region, where it boast­ed in com­pa­ny lit­er­a­ture that it “suc­ceed­ed in main­tain­ing the cohe­sion of the coali­tion to ensure a hard fought vic­to­ry.”

But SCL has also played an equal­ly unset­tling role as a pri­vate arm of British and US counter-insur­gency efforts in the Mid­dle East.

Inter­nal doc­u­ments legal­ly obtained by the Gray­zone Project pro­vide an exclu­sive look at one such effort over­seen by SCL. The mate­ri­als show how the com­pa­ny used psy­cho­log­i­cal pro­fil­ing, “strate­gic com­mu­ni­ca­tions cam­paigns,” and for­eign oper­a­tives, in meth­ods of sur­veil­lance and manip­u­la­tion that par­al­lel the tac­tics that were sub­se­quent­ly used to influ­ence West­ern elec­tions.

A media pro­fes­sion­al whom SCL attempt­ed to recruit for an influ­ence oper­a­tion in Iran described to the Gray­zone Project an array of covert cam­paigns across the region. Speak­ing on the con­di­tion of anonymi­ty, the source said they were solicit­ed in 2009 by a SCL staffer seek­ing a per­son to infil­trate Iran under jour­nal­is­tic cov­er and gath­er data on its pop­u­la­tion.

Though the media pro­fes­sion­al reject­ed the job, express­ing deep reser­va­tions about the company’s empha­sis on sub­terfuge, they described sim­i­lar oper­a­tions they learned about that tar­get­ed pop­u­la­tions in Libya, Pak­istan, and Syr­ia.

The source was told that the Syr­i­an oper­a­tion gath­ered human intel­li­gence by using for­eign­ers either pos­ing as Ara­bic lan­guage stu­dents or enrolled in study abroad-style pro­grams.

SCL has acknowl­edged in com­pa­ny doc­u­ments that it has oper­at­ed in Libya, Syr­ia and Iran.

The doc­u­ments obtained by the Gray­zone Project pro­vide per­haps the first inside look at one of these pro­grams. They shine light on the “research and analy­sis study” that SCL con­duct­ed in Yemen on behalf of the mil­i­tary con­trac­tor Archimedes, known as Project Tita­nia.

‘A Cov­er Sto­ry Will Be Used’

When SCL launched Project Tita­nia, US drones reg­u­lar­ly shad­owed the skies over the Yemeni regions of Marib and Hadra­mout. Al Qae­da in the Ara­bi­an Penin­su­la (AQAP) had suc­cess­ful­ly turned both areas into train­ing and recruit­ment grounds, as well as bases for attacks on gov­ern­ment troops.

In May 2010, an Amer­i­can drone strike acci­den­tal­ly killed the deputy gov­er­nor of Marib, pro­vok­ing his tribe to attack the country’s main oil pipeline in revenge and drain its econ­o­my of $1 bil­lion in rev­enue. A year lat­er in Marib, a drone strike per­son­al­ly autho­rized by then-Pres­i­dent Barack Oba­ma killed AQAP pro­pa­gan­dist Anwar al-Awla­ki and his son, Abdu­rah­man, who was a US cit­i­zen.

Project Tita­nia was defined by its authors as a counter-rad­i­cal­iza­tion study aimed at reduc­ing the appeal and influ­ence of AQAP across Yemen. The oper­a­tion first aimed to iden­ti­fy a “Non Desired Behav­ior” — defined as “sup­port for, and engage­ment in, vio­lent Jihadism” — and to explain the fac­tors that account­ed for its exis­tence and growth. Next, the project called for a pro­pos­al for a “Strate­gic Com­mu­ni­ca­tion Cam­paign” to under­mine the spread of jihadist ide­ol­o­gy.

SCL instruct­ed its field researchers to dis­sem­i­nate ques­tion­naires to Yemeni locals in order to gath­er their psy­cho­log­i­cal pro­files. The goal spelled out in Project Titania’s mis­sion plan was to gain a “detailed psy­choso­cial under­stand­ing of the groups that can be used to most effec­tive­ly influ­ence” young Yemeni men deemed vul­ner­a­ble to jihadist recruit­ment. These meth­ods pre­dat­ed the use by Cam­bridge Ana­lyt­i­ca of Face­book ques­tion­naires to gath­er per­son­al data on Amer­i­can vot­ers.

The for­eign­ers dis­patched to imple­ment Project Tita­nia were explic­it­ly instruct­ed to deceive the Yemeni cit­i­zens they would rely on for field research. “A cov­er sto­ry will be used to explain the pur­pose of the research to the researchers; the local researchers will not be informed of the objec­tives or spon­sors of the study,” a SCL doc­u­ment states. “The ques­tion­naire and inter­view pro­to­col will be com­plete­ly non-attrib­ut­able to the orig­i­nal source.”

SCL even pro­posed options for the cov­er sto­ry: “Pri­or to com­plet­ing the inter­view or the ques­tion­naire, all par­tic­i­pants will be giv­en a ratio­nale for the study (i.e., that the study is part of a uni­ver­si­ty research pro­gramme or a mar­ket research pro­gramme) and they will be informed that their respons­es will be kept con­fi­den­tial.”

The use of cov­er sto­ries in West­ern intel­li­gence oper­a­tions has led to severe social dam­age in some cas­es. In Pak­istan, locals had long sus­pect­ed that the CIA was hunt­ing for drone tar­gets behind the smoke­screen of a vac­ci­na­tion pro­gram run by Pak­istani Polio Erad­i­ca­tion Ini­tia­tive. When news broke that the CIA had run a bogus hepati­tis B erad­i­ca­tion cam­paign in an unsuc­cess­ful bid to obtain the DNA of Osama bin Laden’s fam­i­ly, mil­i­tant ele­ments ini­ti­at­ed a boy­cott of vac­ci­na­tion pro­grams. Over 3 mil­lion chil­dren went with­out polio vac­cines as a result, and the dis­ease spread into neigh­bor­ing coun­tries.

Did the UK Gov­ern­ment Con­tract SCL in Yemen?

The leaked Project Tita­nia doc­u­ments high­light the British government’s par­tic­i­pa­tion in SCL’s covert activ­i­ties in Yemen, and sug­gest that it was the client that had con­tract­ed the pri­vate counter-insur­gency oper­a­tion.

A sec­tion labeled “risk reg­is­ter” con­tains the fol­low­ing secu­ri­ty guide­line for field oper­a­tives: “All for­eign nation­al team mem­bers to liaise with con­tact at British embassy and reg­is­ter with UK FCO LOCATE ser­vice.” (The British For­eign and Com­mon­wealth Office’s LOCATE ser­vice pro­vid­ed expats with secu­ri­ty issue updates; it was abol­ished in 2013 because so few cit­i­zens enrolled in it).

Project Tita­nia doc­u­ments list­ed an array of NGOs oper­at­ing in Marib and appeared to sug­gest them as poten­tial vehi­cles for obtain­ing intel­li­gence on the indige­nous pop­u­la­tion. The names of those NGOs have been redact­ed from this report to pro­tect staff from reper­cus­sions that might result from their con­nec­tion to a covert West­ern influ­ence oper­a­tion. If any wound up as par­tic­i­pants in Project Tita­nia, they did so unwit­ting­ly, as project mate­ri­als specif­i­cal­ly demand­ed they be coerced into the oper­a­tion under false pre­tens­es.

Yasha Levine, a jour­nal­ist and author of “Sur­veil­lance Val­ley: The Hid­den His­to­ry of the Inter­net,” saw SCL’s Project Tita­nia as a fair­ly typ­i­cal counter-insur­gency oper­a­tion. “Look­ing at these doc­u­ments about SCL’s Yemen data-dri­ven coun­terin­sur­gency pro­gram, the most remark­able thing about them is just how unre­mark­able it is,” Levine told the Gray­zone Project.

“If you change the word­ing a bit,” he con­tin­ued, “SCL’s pro­pos­al could have been writ­ten a half cen­tu­ry ago for the Viet­nam War, where com­put­er-aid­ed coun­terin­sur­gency tech­nolo­gies were first pio­neered. Back then the idea was that to fight insur­gen­cies — wars in which ene­my com­bat­ants came out of the gen­er­al pop­u­la­tion — you first need­ed to under­stand the cul­tur­al, social and polit­i­cal envi­ron­ment in which the ene­my oper­at­ed. That meant first and fore­most study­ing and sur­veilling restive pop­u­la­tions as if they were lab rats, and then using advanced com­put­er tech­nol­o­gy to shift through and process all the infor­ma­tion com­ing in.”

...

But there was more to Project Tita­nia than what was revealed in the SCL doc­u­ments. Com­mu­ni­ca­tions obtained by the Gray­zone Project linked the project’s prin­ci­pal direc­tor to a much wider pro­gram of sur­veil­lance and data min­ing aimed not only at a region or two, but at the entire Arab world.

———-

“Exclu­sive Leaked Docs Expose Yemen-Based Counter-Insur­gency Pro­gram by Cam­bridge Ana­lyt­i­ca Par­ent Com­pa­ny SCL” by Max Blu­men­thal; The Gray­zone Project; 05/23/2018

“Found­ed in 2005, SCL spe­cial­izes in what com­pa­ny lit­er­a­ture has described as “influ­ence oper­a­tions” and “psy­cho­log­i­cal war­fare” around the globe. An SCL brochure leaked to the BBC revealed how the firm exac­er­bat­ed eth­nic ten­sions in Latvia to assist their client in 2006.”

SCL’s found­ing doc­u­ments going back to 2005 tout its abil­i­ty to wage “influ­ence oper­a­tions” and “psy­cho­log­i­cal war­fare” around the globe. That’s how far back the Cam­bridge Ana­lyt­i­ca sto­ry goes. Although it appears to go even fur­ther back since SCL’s brochure boast­ed of its suc­cess “in main­tain­ing the cohe­sion of the coali­tion to ensure a hard fought vic­to­ry,” of the 2004 Orange Rev­o­lu­tion in Ukraine:

...
A year lat­er, the firm orches­trat­ed “anti-elec­tion” ral­lies to sup­press the oppo­si­tion vote in Nigeria’s 2007 pres­i­den­tial elec­tion. Then, in 2010, accord­ing to the BBC, SCL ini­ti­at­ed an “ambi­tious cam­paign of polit­i­cal graf­fi­ti” that “osten­si­bly came from the youth,” enabling its client to “claim cred­it for lis­ten­ing to a ‘unit­ed youth.’”

SCL has also applied its influ­ence in Ukraine, first as part of the broad­er pub­lic rela­tions cam­paign dur­ing the country’s NATO-backed 2004 “Orange Rev­o­lu­tion,” then in Ukraine’s con­test­ed Don­bas region, where it boast­ed in com­pa­ny lit­er­a­ture that it “suc­ceed­ed in main­tain­ing the cohe­sion of the coali­tion to ensure a hard fought vic­to­ry.”
...

Lat­er, in 2009, SCL was doing some sort of psy­cho­log­i­cal pro­fil­ing Iran. Along with Libya, Pak­istan, and Syr­ia:

...
But SCL has also played an equal­ly unset­tling role as a pri­vate arm of British and US counter-insur­gency efforts in the Mid­dle East.

Inter­nal doc­u­ments legal­ly obtained by the Gray­zone Project pro­vide an exclu­sive look at one such effort over­seen by SCL. The mate­ri­als show how the com­pa­ny used psy­cho­log­i­cal pro­fil­ing, “strate­gic com­mu­ni­ca­tions cam­paigns,” and for­eign oper­a­tives, in meth­ods of sur­veil­lance and manip­u­la­tion that par­al­lel the tac­tics that were sub­se­quent­ly used to influ­ence West­ern elec­tions.

A media pro­fes­sion­al whom SCL attempt­ed to recruit for an influ­ence oper­a­tion in Iran described to the Gray­zone Project an array of covert cam­paigns across the region. Speak­ing on the con­di­tion of anonymi­ty, the source said they were solicit­ed in 2009 by a SCL staffer seek­ing a per­son to infil­trate Iran under jour­nal­is­tic cov­er and gath­er data on its pop­u­la­tion.

Though the media pro­fes­sion­al reject­ed the job, express­ing deep reser­va­tions about the company’s empha­sis on sub­terfuge, they described sim­i­lar oper­a­tions they learned about that tar­get­ed pop­u­la­tions in Libya, Pak­istan, and Syr­ia.

The source was told that the Syr­i­an oper­a­tion gath­ered human intel­li­gence by using for­eign­ers either pos­ing as Ara­bic lan­guage stu­dents or enrolled in study abroad-style pro­grams.

SCL has acknowl­edged in com­pa­ny doc­u­ments that it has oper­at­ed in Libya, Syr­ia and Iran.
...

So years before the 2016 elec­tion, SCL was already act­ing as a psy­cho­log­i­cal war­fare con­trac­tor in coun­tries around the world. It points to anoth­er impor­tant con­text for the Cam­bridge Ana­lyt­i­ca scan­dal: the US pop­u­lace tar­get­ed in 2016 may have effec­tive­ly been guinea pigs for this tech­nol­o­gy in the con­text of using Face­book to gath­er psy­cho­log­i­cal pro­files on large num­bers of peo­ple. But they weren’t the first guinea pigs on SCL’s psy­cho­log­i­cal pro­fil­ing tech­niques because that’s what SCL has been for years in soci­eties across the world. Appar­ent­ly start­ing in Ukraine.

So this sto­ry is promis­ing to get much big­ger as more doc­u­ments are leaked. It also rais­es an inter­est­ing ques­tion in the con­text of Pres­i­dent Trump’s deci­sion to drone assas­si­nate one of Iran’s most revered lead­ers: from a psy­cho­log­i­cal war­fare per­spec­tive, was that a good idea? It does­n’t seem like it was a very good idea, but it would be inter­est­ing to know what the regime change psy­cho­log­i­cal war­fare spe­cial­ists say about that. Since Cam­bridge Ana­lyt­i­ca has unfor­tu­nate­ly rein­cor­po­rat­ed as Emer­da­ta maybe some­one can ask them about that.

The New York Times had a recent piece about a com­pa­ny that’s described as a lit­tle-known enti­ty that might end pri­va­cy as we know it. Basi­cal­ly, the com­pa­ny, Clearview AI, offers what amounts to a super-facial recog­ni­tion ser­vice. The com­pa­ny appears to have scraped as much image and iden­ti­ty infor­ma­tion as pos­si­ble from social media sites like Face­book, YouTube, and Ven­mo and allows clients to upload a pic­ture of any­one and see per­son­al pro­files on all of those match­es. Those pro­files include all of the match­ing pic­tures as well as links to where those pic­tures appeared. So it’s like a search­able data­base of bil­lions of pho­tos and ids, where you start the search with a pho­to and it returns more pho­tos and infor­ma­tion on every­one who is a close enough match. The data­base of more than 3 bil­lion pic­tures is described as being far beyond any­thing ever con­struct­ed by the US gov­ern­ment or Sil­i­con Val­ley giants. In addi­tion, Clearview is devel­op­ing a pair of glass­es that will give the wear­er a heads-up dis­play of the names and infor­ma­tion of any­one you’re look­ing at in real-time.

And while the com­pa­ny is appar­ent­ly quite tiny and lit­tle known to the pub­lic, it’s ser­vices have already been used by over 600 law enforce­ment agen­cies in the US. But it’s not just law enforce­ment using these ser­vices. We’re also told the soft­ware has been licensed to pri­vate com­pa­nies for secu­ri­ty pur­pos­es, although we aren’t told the names of those com­pa­nies.

All in all, it’s a pret­ty trou­bling com­pa­ny. But it of course gets much worse. It turns out the com­pa­ny is heav­i­ly con­nect­ed to the Repub­li­can Par­ty and large­ly rely­ing on Repub­li­cans to pro­mote it to poten­tial clients. The com­pa­ny was co-found­ed by an Aus­tralian, Hoan Ton-That, and Richard Schwartz. Ton-That worked on devel­op­ing the ini­tial tech­nol­o­gy and Schwartz was respon­si­ble for lin­ing up poten­tial clients. Schwartz is the long-time senior aide to Rudy Giu­liani and has quite an exten­sive Rolodex. Schwartz report­ed­ly met Ton-That in 2016 at a book event at the con­ser­v­a­tive Man­hat­tan Insti­tute. So it sounds like Ton-That was already work­ing on net­work­ing with­in right-wing cir­cles when he met Schwarz.

By the end of 2017, the com­pa­ny had its facial recog­ni­tion project ready to start pitch­ing to clients. The way Ton-That describes it, they were try­ing to think of any pos­si­ble client who might be inter­est­ed in this tech­nol­o­gy, like par­ents who want to do a back­ground check on a poten­tial baby-sit­ter or an add-on fea­ture for secu­ri­ty cam­eras. In oth­er words, they have plans on even­tu­al­ly releas­ing this tech­nol­o­gy to any­one.

One of the peo­ple they made their ini­tial pitch to was Paul Nehlen, the for­mer Repub­li­can ris­ing star who ran for Paul Ryan’s for­mer House seat but even­tu­al­ly out­ed him­self as a vir­u­lent neo-Nazi. Clearview was offer­ing their ser­vices to Nehlen dur­ing his cam­paign for “extreme oppo­si­tion research”. In oth­er words, they were pre­sum­ably going to use the data­base to find all visu­al records of Nehlen’s oppo­nents and the peo­ple work­ing for their cam­paign to dig up dirt. So this com­pa­ny is start­ed by a bunch of Repub­li­cans and one of the first client pitch­es they make is to a neo-Nazi Repub­li­can. It gives us a sense of the pol­i­tics of this com­pa­ny.

The failed pitch to Nehlen was made in late 2017 and we’re told that soon after that the com­pa­ny got its first round of fund­ing from out­side investors. One of those investors was Peter Thiel, who made a $200,000 invest­ment. Accord­ing to Thiel’s spokesman, “In 2017, Peter gave a tal­ent­ed young founder $200,000, which two years lat­er con­vert­ed to equi­ty in Clearview AI,” and, “That was Peter’s only con­tri­bu­tion; he is not involved in the com­pa­ny.” So Thiel made one of the first invest­ments which was con­vert­ed to equi­ty, mean­ing he’s a share­hold­er now. But we’re told he’s not involved in the com­pa­ny, which sounds like a typ­i­cal Thiel decep­tion.

Keep in mind that Thiel is in a posi­tion to both encour­age the hand­ing of large vol­umes of faces and IDs to the com­pa­ny while also being a posi­tion to mas­sive­ly exploit Clearview’s tech­nol­o­gy. Thiel co-found­ed Palan­tir, which could obvi­ous­ly have exten­sive uses for this tech­nol­o­gy, and Thiel also sits on the board of Face­book, where much of the pho­tos and ID infor­ma­tion was scraped. When asked about Clearview’s scrap­ing of Face­book data to pop­u­lat­ed its data­base, Face­book said the com­pa­ny if review­ing the sit­u­a­tion and “will take appro­pri­ate action if we find they are vio­lat­ing our rules.” But Face­book had no com­ment on the fact that Thiel sits on its board and is per­son­al­ly invest­ed in Clearview. Accord­ing to Ton-That, “A lot of peo­ple are doing it,” and, “Face­book knows.”

Oth­er Repub­li­can Par­ty con­nec­tions to the com­pa­ny include Jes­si­ca Medeiros Gar­ri­son and Bran­don Fricke. Medeiros Gar­ris­son, the main con­tact for cus­tomers, man­aged Luther Strange’s Repub­li­can cam­paign for Alaba­ma attor­ney gen­er­al while Fricke, a “growth con­sul­tant” for the com­pa­ny is engaged to right-wing media per­son­al­i­ty Tomi Lahren. Clearview claims it’s also enlist­ed Democ­rats to mar­ket its prod­ucts too but we aren’t giv­en any names of those Democ­rats.

So how does Clearview assuage con­cerns about the legal­i­ty of its ser­vices? That job falls to Paul D. Clement, a Unit­ed States solic­i­tor gen­er­al under Pres­i­dent George W. Bush. Paul Clement, a for­mer clerk to Antonin Scalia, has the inter­est­ing dis­tinc­tion for a Repub­li­can lawyer. In 2012, Clement was the lawyer who led the Repub­li­can chal­lenge by 26 states in 2012 to repeal Oba­macare over its indi­vid­ual man­date pro­vi­sion. That’s some­thing we would expect for a for­mer Bush admin­is­tra­tion offi­cial. But back in Octo­ber of 2019, Clement was asked by the Supreme Court to defend an Oba­ma-era law after the inde­pen­dence of the head of the Con­sumer Finan­cial Pro­tec­tion Bureau (CFPB) was chal­lenged by the Trump admin­is­tra­tion’s Jus­tice Depart­ment. The CFPB itself (which is now head­ed by a Trump appointee) also joined the Jus­tice Depart­ment in the law­suit, leav­ing no enti­ty to defend the orig­i­nal law that pre­vents pres­i­dents from fir­ing the heads of the CFPB. The CFPB was one of the enti­ties set up by the Oba­ma admin­is­tra­tion (and designed by Sen­a­tor Eliz­a­beth War­ren) fol­low­ing the finan­cial cri­sis so it was guar­an­teed the Trump admin­is­tra­tion would oppose it. The fact that it’s ded­i­cat­ed to pro­vid­ing con­sumer finan­cial pro­tec­tion is the oth­er rea­son it was guar­an­teed the Trump admin­is­tra­tion would opposed it. Repub­li­cans don’t do con­sumer pro­tec­tion. The Oba­ma-appoint­ed head of the CFPB, Richard Cor­dray, resigned in Novem­ber of 2017 two years ear­ly after Trump and the Repub­li­cans made it abun­dant­ly clear they want­ed to replace him. The Trump Jus­tice Depart­ment argued back in March of 2017 that this restric­tion on the pres­i­den­t’s abil­i­ty to fire the head of the CFPB made it uncon­sti­tu­tion­al. In Sep­tem­ber of 2019, the Jus­tice Depart­ment as the Supreme Court to take the case, and the fol­low­ing month Clement — who has argued before the Supreme Court more than 95 times — was invit­ed by the Supreme Court to defend the exist­ing struc­ture of the CFPB.

Oh, and Paul D. Clement also hap­pens to be one of the lawyers who suc­cess­ful­ly argued on behalf of the Repub­li­cans in Rucho v. Com­mon Cause, a case that has now con­sti­tu­tion­al­ly enshrined hyper-par­ti­san ger­ry­man­der­ing that the fed­er­al courts can do noth­ing about. So that gives us a sense of the impor­tance of hav­ing some­how like Paul D. Clements solitic­ing clients for a com­pa­ny like ClearView: while he’s an extreme­ly high pro­file and respect­ed lawyer, he’s also a par­ti­san hack. But the kind of hack whose words will car­ry a lot of weight when it comes to assur­ing poten­tial clients about the legal­i­ty of Clearview’s prod­ucts.

And if that all was­n’t shady enough, the author of the fol­low­ing report shares an anec­dote that should raise big red flags about the char­ac­ter of the peo­ple behind this com­pa­ny: When the author test­ed the sys­tem on his own pho­to by ask­ing a friend in law enforce­ment to run his pic­ture through it, the jour­nal­ist got dozens of pic­tures of him­self back includ­ing some pic­tures he did­n’t even know exist­ed. But his law enforce­ment friend was soon con­tact­ed by Clearview to ask if he had been speak­ing to the media. So Clearview is either active­ly mon­i­tor­ing and doing its own search­es on the peo­ple run through its sys­tem or it has sys­tem set up to flag ‘trou­ble­mak­ers’ like jour­nal­ists. Ton-That claims that the rea­son this search prompt­ed a call from the com­pa­ny is because the sys­tem is set up to flag “pos­si­ble anom­alous search behav­ior” in order to pre­vent “inap­pro­pri­ate search­es.” But after that inci­dent, the report­ed found that his results were removed from future search­es, which Ton-That dis­missed as a “soft­ware bug”. So the com­pa­ny appears to be active­ly mon­i­tor­ing and manip­u­lat­ing search results. As the arti­cle notes, since the pri­ma­ry users of Clearview are police agen­cies at this point, the com­pa­ny can get a detailed list of peo­ple who have received the inter­est of law enforce­ment sim­ply by look­ing at the search­es used, which is the kind of infor­ma­tion that can be poten­tial­ly abused. It’s an exam­ple of why the char­ac­ter of the peo­ple behind this firm is par­tic­u­lar­ly impor­tant for a firm offer­ing these kinds of ser­vices and why the more we’re learn­ing about this com­pa­ny the more cause there is for seri­ous con­cern.

It remains unclear how many clients out­side of law enforce­ment will be allowed to pur­chase Clearview’s ser­vices. But as the arti­cle notes, now that Clearview has bro­ken the taboo of offer­ing facial recog­ni­tion soft­ware data­base ser­vices like this, it’s just a mat­ter of time before com­pa­nies do the same thing. And that’s why Clearview might end up end­ing pri­va­cy as we know it: by set­ting a real­ly, real­ly bad exam­ple by show­ing the world this ser­vice is pos­si­ble and there’s a mar­ket for it:

The New York Times

The Secre­tive Com­pa­ny That Might End Pri­va­cy as We Know It
A lit­tle-known start-up helps law enforce­ment match pho­tos of unknown peo­ple to their online images — and “might lead to a dystopi­an future or some­thing,” a backer says.

By Kash­mir Hill
Jan. 18, 2020

Until recent­ly, Hoan Ton-That’s great­est hits includ­ed an obscure iPhone game and an app that let peo­ple put Don­ald Trump’s dis­tinc­tive yel­low hair on their own pho­tos.

Then Mr. Ton-That — an Aus­tralian techie and one­time mod­el — did some­thing momen­tous: He invent­ed a tool that could end your abil­i­ty to walk down the street anony­mous­ly, and pro­vid­ed it to hun­dreds of law enforce­ment agen­cies, rang­ing from local cops in Flori­da to the F.B.I. and the Depart­ment of Home­land Secu­ri­ty.

His tiny com­pa­ny, Clearview AI, devised a ground­break­ing facial recog­ni­tion app. You take a pic­ture of a per­son, upload it and get to see pub­lic pho­tos of that per­son, along with links to where those pho­tos appeared. The sys­tem — whose back­bone is a data­base of more than three bil­lion images that Clearview claims to have scraped from Face­book, YouTube, Ven­mo and mil­lions of oth­er web­sites — goes far beyond any­thing ever con­struct­ed by the Unit­ed States gov­ern­ment or Sil­i­con Val­ley giants.

Fed­er­al and state law enforce­ment offi­cers said that while they had only lim­it­ed knowl­edge of how Clearview works and who is behind it, they had used its app to help solve shoplift­ing, iden­ti­ty theft, cred­it card fraud, mur­der and child sex­u­al exploita­tion cas­es.

Until now, tech­nol­o­gy that read­i­ly iden­ti­fies every­one based on his or her face has been taboo because of its rad­i­cal ero­sion of pri­va­cy. Tech com­pa­nies capa­ble of releas­ing such a tool have refrained from doing so; in 2011, Google’s chair­man at the time said it was the one tech­nol­o­gy the com­pa­ny had held back because it could be used “in a very bad way.” Some large cities, includ­ing San Fran­cis­co, have barred police from using facial recog­ni­tion tech­nol­o­gy.

But with­out pub­lic scruti­ny, more than 600 law enforce­ment agen­cies have start­ed using Clearview in the past year, accord­ing to the com­pa­ny, which declined to pro­vide a list. The com­put­er code under­ly­ing its app, ana­lyzed by The New York Times, includes pro­gram­ming lan­guage to pair it with aug­ment­ed-real­i­ty glass­es; users would poten­tial­ly be able to iden­ti­fy every per­son they saw. The tool could iden­ti­fy activists at a protest or an attrac­tive stranger on the sub­way, reveal­ing not just their names but where they lived, what they did and whom they knew.

And it’s not just law enforce­ment: Clearview has also licensed the app to at least a hand­ful of com­pa­nies for secu­ri­ty pur­pos­es.

“The weaponiza­tion pos­si­bil­i­ties of this are end­less,” said Eric Gold­man, co-direc­tor of the High Tech Law Insti­tute at San­ta Clara Uni­ver­si­ty. “Imag­ine a rogue law enforce­ment offi­cer who wants to stalk poten­tial roman­tic part­ners, or a for­eign gov­ern­ment using this to dig up secrets about peo­ple to black­mail them or throw them in jail.”

Clearview has shroud­ed itself in secre­cy, avoid­ing debate about its bound­ary-push­ing tech­nol­o­gy. When I began look­ing into the com­pa­ny in Novem­ber, its web­site was a bare page show­ing a nonex­is­tent Man­hat­tan address as its place of busi­ness. The company’s one employ­ee list­ed on LinkedIn, a sales man­ag­er named “John Good,” turned out to be Mr. Ton-That, using a fake name. For a month, peo­ple affil­i­at­ed with the com­pa­ny would not return my emails or phone calls.

While the com­pa­ny was dodg­ing me, it was also mon­i­tor­ing me. At my request, a num­ber of police offi­cers had run my pho­to through the Clearview app. They soon received phone calls from com­pa­ny rep­re­sen­ta­tives ask­ing if they were talk­ing to the media — a sign that Clearview has the abil­i­ty and, in this case, the appetite to mon­i­tor whom law enforce­ment is search­ing for.

Facial recog­ni­tion tech­nol­o­gy has always been con­tro­ver­sial. It makes peo­ple ner­vous about Big Broth­er. It has a ten­den­cy to deliv­er false match­es for cer­tain groups, like peo­ple of col­or. And some facial recog­ni­tion prod­ucts used by the police — includ­ing Clearview’s — haven’t been vet­ted by inde­pen­dent experts.

Clearview’s app car­ries extra risks because law enforce­ment agen­cies are upload­ing sen­si­tive pho­tos to the servers of a com­pa­ny whose abil­i­ty to pro­tect its data is untest­ed.

The com­pa­ny even­tu­al­ly start­ed answer­ing my ques­tions, say­ing that its ear­li­er silence was typ­i­cal of an ear­ly-stage start-up in stealth mode. Mr. Ton-That acknowl­edged design­ing a pro­to­type for use with aug­ment­ed-real­i­ty glass­es but said the com­pa­ny had no plans to release it. And he said my pho­to had rung alarm bells because the app “flags pos­si­ble anom­alous search behav­ior” in order to pre­vent users from con­duct­ing what it deemed “inap­pro­pri­ate search­es.”

In addi­tion to Mr. Ton-That, Clearview was found­ed by Richard Schwartz — who was an aide to Rudolph W. Giu­liani when he was may­or of New York — and backed finan­cial­ly by Peter Thiel, a ven­ture cap­i­tal­ist behind Face­book and Palan­tir.

Anoth­er ear­ly investor is a small firm called Kire­na­ga Part­ners. Its founder, David Scal­zo, dis­missed con­cerns about Clearview mak­ing the inter­net search­able by face, say­ing it’s a valu­able crime-solv­ing tool.

“I’ve come to the con­clu­sion that because infor­ma­tion con­stant­ly increas­es, there’s nev­er going to be pri­va­cy,” Mr. Scal­zo said. “Laws have to deter­mine what’s legal, but you can’t ban tech­nol­o­gy. Sure, that might lead to a dystopi­an future or some­thing, but you can’t ban it.”

Addict­ed to A.I.

Mr. Ton-That, 31, grew up a long way from Sil­i­con Val­ley. In his native Aus­tralia, he was raised on tales of his roy­al ances­tors in Viet­nam. In 2007, he dropped out of col­lege and moved to San Fran­cis­co. The iPhone had just arrived, and his goal was to get in ear­ly on what he expect­ed would be a vibrant mar­ket for social media apps. But his ear­ly ven­tures nev­er gained real trac­tion.

In 2009, Mr. Ton-That cre­at­ed a site that let peo­ple share links to videos with all the con­tacts in their instant mes­sen­gers. Mr. Ton-That shut it down after it was brand­ed a “phish­ing scam.” In 2015, he spun up Trump Hair, which added Mr. Trump’s dis­tinc­tive coif to peo­ple in a pho­to, and a pho­to-shar­ing pro­gram. Both fiz­zled.

Dispir­it­ed, Mr. Ton-That moved to New York in 2016. Tall and slen­der, with long black hair, he con­sid­ered a mod­el­ing career, he said, but after one shoot he returned to try­ing to fig­ure out the next big thing in tech. He start­ed read­ing aca­d­e­m­ic papers on arti­fi­cial intel­li­gence, image recog­ni­tion and machine learn­ing.

Mr. Schwartz and Mr. Ton-That met in 2016 at a book event at the Man­hat­tan Insti­tute, a con­ser­v­a­tive think tank. Mr. Schwartz, now 61, had amassed an impres­sive Rolodex work­ing for Mr. Giu­liani in the 1990s and serv­ing as the edi­to­r­i­al page edi­tor of The New York Dai­ly News in the ear­ly 2000s. The two soon decid­ed to go into the facial recog­ni­tion busi­ness togeth­er: Mr. Ton-That would build the app, and Mr. Schwartz would use his con­tacts to drum up com­mer­cial inter­est.

Police depart­ments have had access to facial recog­ni­tion tools for almost 20 years, but they have his­tor­i­cal­ly been lim­it­ed to search­ing gov­ern­ment-pro­vid­ed images, such as mug shots and driver’s license pho­tos. In recent years, facial recog­ni­tion algo­rithms have improved in accu­ra­cy, and com­pa­nies like Ama­zon offer prod­ucts that can cre­ate a facial recog­ni­tion pro­gram for any data­base of images.

Mr. Ton-That want­ed to go way beyond that. He began in 2016 by recruit­ing a cou­ple of engi­neers. One helped design a pro­gram that can auto­mat­i­cal­ly col­lect images of people’s faces from across the inter­net, such as employ­ment sites, news sites, edu­ca­tion­al sites, and social net­works includ­ing Face­book, YouTube, Twit­ter, Insta­gram and even Ven­mo. Rep­re­sen­ta­tives of those com­pa­nies said their poli­cies pro­hib­it such scrap­ing, and Twit­ter said it explic­it­ly banned use of its data for facial recog­ni­tion.

Anoth­er engi­neer was hired to per­fect a facial recog­ni­tion algo­rithm that was derived from aca­d­e­m­ic papers. The result: a sys­tem that uses what Mr. Ton-That described as a “state-of-the-art neur­al net” to con­vert all the images into math­e­mat­i­cal for­mu­las, or vec­tors, based on facial geom­e­try — like how far apart a person’s eyes are. Clearview cre­at­ed a vast direc­to­ry that clus­tered all the pho­tos with sim­i­lar vec­tors into “neigh­bor­hoods.” When a user uploads a pho­to of a face into Clearview’s sys­tem, it con­verts the face into a vec­tor and then shows all the scraped pho­tos stored in that vector’s neigh­bor­hood — along with the links to the sites from which those images came.

Mr. Schwartz paid for serv­er costs and basic expens­es, but the oper­a­tion was bare bones; every­one worked from home. “I was liv­ing on cred­it card debt,” Mr. Ton-That said. “Plus, I was a Bit­coin believ­er, so I had some of those.”

Going Viral With Law Enforce­ment

By the end of 2017, the com­pa­ny had a for­mi­da­ble facial recog­ni­tion tool, which it called Smartcheckr. But Mr. Schwartz and Mr. Ton-That weren’t sure whom they were going to sell it to.

Maybe it could be used to vet babysit­ters or as an add-on fea­ture for sur­veil­lance cam­eras. What about a tool for secu­ri­ty guards in the lob­bies of build­ings or to help hotels greet guests by name? “We thought of every idea,” Mr. Ton-That said.

One of the odd­er pitch­es, in late 2017, was to Paul Nehlen — an anti-Semi­te and self-described “pro-white” Repub­li­can run­ning for Con­gress in Wis­con­sin — to use “uncon­ven­tion­al data­bas­es” for “extreme oppo­si­tion research,” accord­ing to a doc­u­ment pro­vid­ed to Mr. Nehlen and lat­er post­ed online. Mr. Ton-That said the com­pa­ny nev­er actu­al­ly offered such ser­vices.

The com­pa­ny soon changed its name to Clearview AI and began mar­ket­ing to law enforce­ment. That was when the com­pa­ny got its first round of fund­ing from out­side investors: Mr. Thiel and Kire­na­ga Part­ners. Among oth­er things, Mr. Thiel was famous for secret­ly financ­ing Hulk Hogan’s law­suit that bank­rupt­ed the pop­u­lar web­site Gawk­er. Both Mr. Thiel and Mr. Ton-That had been the sub­ject of neg­a­tive arti­cles by Gawk­er.

“In 2017, Peter gave a tal­ent­ed young founder $200,000, which two years lat­er con­vert­ed to equi­ty in Clearview AI,” said Jere­mi­ah Hall, Mr. Thiel’s spokesman. “That was Peter’s only con­tri­bu­tion; he is not involved in the com­pa­ny.”

Even after a sec­ond fund­ing round in 2019, Clearview remains tiny, hav­ing raised $7 mil­lion from investors, accord­ing to Pitch­book, a web­site that tracks invest­ments in start-ups. The com­pa­ny declined to con­firm the amount.

In Feb­ru­ary, the Indi­ana State Police start­ed exper­i­ment­ing with Clearview. They solved a case with­in 20 min­utes of using the app. Two men had got­ten into a fight in a park, and it end­ed when one shot the oth­er in the stom­ach. A bystander record­ed the crime on a phone, so the police had a still of the gunman’s face to run through Clearview’s app.

They imme­di­ate­ly got a match: The man appeared in a video that some­one had post­ed on social media, and his name was includ­ed in a cap­tion on the video. “He did not have a driver’s license and hadn’t been arrest­ed as an adult, so he wasn’t in gov­ern­ment data­bas­es,” said Chuck Cohen, an Indi­ana State Police cap­tain at the time.

The man was arrest­ed and charged; Mr. Cohen said he prob­a­bly wouldn’t have been iden­ti­fied with­out the abil­i­ty to search social media for his face. The Indi­ana State Police became Clearview’s first pay­ing cus­tomer, accord­ing to the com­pa­ny. (The police declined to com­ment beyond say­ing that they test­ed Clearview’s app.)

Clearview deployed cur­rent and for­mer Repub­li­can offi­cials to approach police forces, offer­ing free tri­als and annu­al licens­es for as lit­tle as $2,000. Mr. Schwartz tapped his polit­i­cal con­nec­tions to help make gov­ern­ment offi­cials aware of the tool, accord­ing to Mr. Ton-That. (“I’m thrilled to have the oppor­tu­ni­ty to help Hoan build Clearview into a mis­sion-dri­ven orga­ni­za­tion that’s help­ing law enforce­ment pro­tect chil­dren and enhance the safe­ty of com­mu­ni­ties across the coun­try,” Mr. Schwartz said through a spokes­woman.)

The company’s main con­tact for cus­tomers was Jes­si­ca Medeiros Gar­ri­son, who man­aged Luther Strange’s Repub­li­can cam­paign for Alaba­ma attor­ney gen­er­al. Bran­don Fricke, an N.F.L. agent engaged to the Fox Nation host Tomi Lahren, said in a finan­cial dis­clo­sure report dur­ing a con­gres­sion­al cam­paign in Cal­i­for­nia that he was a “growth con­sul­tant” for the com­pa­ny. (Clearview said that it was a brief, unpaid role, and that the com­pa­ny had enlist­ed Democ­rats to help mar­ket its prod­uct as well.)

The company’s most effec­tive sales tech­nique was offer­ing 30-day free tri­als to offi­cers, who then encour­aged their acqui­si­tion depart­ments to sign up and praised the tool to offi­cers from oth­er police depart­ments at con­fer­ences and online, accord­ing to the com­pa­ny and doc­u­ments pro­vid­ed by police depart­ments in response to pub­lic-record requests. Mr. Ton-That final­ly had his viral hit.

In July, a detec­tive in Clifton, N.J., urged his cap­tain in an email to buy the soft­ware because it was “able to iden­ti­fy a sus­pect in a mat­ter of sec­onds.” Dur­ing the department’s free tri­al, Clearview had iden­ti­fied shoplifters, an Apple Store thief and a good Samar­i­tan who had punched out a man threat­en­ing peo­ple with a knife.

Pho­tos “could be covert­ly tak­en with tele­pho­to lens and input into the soft­ware, with­out ‘burn­ing’ the sur­veil­lance oper­a­tion,” the detec­tive wrote in the email, pro­vid­ed to The Times by two researchers, Beryl Lip­ton of Muck­Rock and Fred­dy Mar­tinez of Open the Gov­ern­ment. They dis­cov­ered Clearview late last year while look­ing into how local police depart­ments are using facial recog­ni­tion.

...

In Gainesville, Fla., Detec­tive Sgt. Nick Fer­rara heard about Clearview last sum­mer when it adver­tised on CrimeDex, a list-serv for inves­ti­ga­tors who spe­cial­ize in finan­cial crimes. He said he had pre­vi­ous­ly relied sole­ly on a state-pro­vid­ed facial recog­ni­tion tool, FACES, which draws from more than 30 mil­lion Flori­da mug shots and Depart­ment of Motor Vehi­cle pho­tos.

Sergeant Fer­rara found Clearview’s app supe­ri­or, he said. Its nation­wide data­base of images is much larg­er, and unlike FACES, Clearview’s algo­rithm doesn’t require pho­tos of peo­ple look­ing straight at the cam­era.

“With Clearview, you can use pho­tos that aren’t per­fect,” Sergeant Fer­rara said. “A per­son can be wear­ing a hat or glass­es, or it can be a pro­file shot or par­tial view of their face.”

He uploaded his own pho­to to the sys­tem, and it brought up his Ven­mo page. He ran pho­tos from old, dead-end cas­es and iden­ti­fied more than 30 sus­pects. In Sep­tem­ber, the Gainesville Police Depart­ment paid $10,000 for an annu­al Clearview license.

Fed­er­al law enforce­ment, includ­ing the F.B.I. and the Depart­ment of Home­land Secu­ri­ty, are try­ing it, as are Cana­di­an law enforce­ment author­i­ties, accord­ing to the com­pa­ny and gov­ern­ment offi­cials.

Despite its grow­ing pop­u­lar­i­ty, Clearview avoid­ed pub­lic men­tion until the end of 2019, when Flori­da pros­e­cu­tors charged a woman with grand theft after two grills and a vac­u­um were stolen from an Ace Hard­ware store in Cler­mont. She was iden­ti­fied when the police ran a still from a sur­veil­lance video through Clearview, which led them to her Face­book page. A tat­too vis­i­ble in the sur­veil­lance video and Face­book pho­tos con­firmed her iden­ti­ty, accord­ing to an affi­davit in the case.

‘We’re All Screwed’

Mr. Ton-That said the tool does not always work. Most of the pho­tos in Clearview’s data­base are tak­en at eye lev­el. Much of the mate­r­i­al that the police upload is from sur­veil­lance cam­eras mount­ed on ceil­ings or high on walls.

“They put sur­veil­lance cam­eras too high,” Mr. Ton-That lament­ed. “The angle is wrong for good face recog­ni­tion.”

Despite that, the com­pa­ny said, its tool finds match­es up to 75 per­cent of the time. But it is unclear how often the tool deliv­ers false match­es, because it has not been test­ed by an inde­pen­dent par­ty such as the Nation­al Insti­tute of Stan­dards and Tech­nol­o­gy, a fed­er­al agency that rates the per­for­mance of facial recog­ni­tion algo­rithms.

“We have no data to sug­gest this tool is accu­rate,” said Clare Garvie, a researcher at George­town University’s Cen­ter on Pri­va­cy and Tech­nol­o­gy, who has stud­ied the government’s use of facial recog­ni­tion. “The larg­er the data­base, the larg­er the risk of misiden­ti­fi­ca­tion because of the dop­pel­gänger effect. They’re talk­ing about a mas­sive data­base of ran­dom peo­ple they’ve found on the inter­net.”

But cur­rent and for­mer law enforce­ment offi­cials say the app is effec­tive. “For us, the test­ing was whether it worked or not,” said Mr. Cohen, the for­mer Indi­ana State Police cap­tain.

One rea­son that Clearview is catch­ing on is that its ser­vice is unique. That’s because Face­book and oth­er social media sites pro­hib­it peo­ple from scrap­ing users’ images — Clearview is vio­lat­ing the sites’ terms of ser­vice.

“A lot of peo­ple are doing it,” Mr. Ton-That shrugged. “Face­book knows.”

Jay Nan­car­row, a Face­book spokesman, said the com­pa­ny was review­ing the sit­u­a­tion with Clearview and “will take appro­pri­ate action if we find they are vio­lat­ing our rules.”

Mr. Thiel, the Clearview investor, sits on Facebook’s board. Mr. Nan­car­row declined to com­ment on Mr. Thiel’s per­son­al invest­ments.

Some law enforce­ment offi­cials said they didn’t real­ize the pho­tos they uploaded were being sent to and stored on Clearview’s servers. Clearview tries to pre-empt con­cerns with an F.A.Q. doc­u­ment giv­en to would-be clients that says its cus­tomer-sup­port employ­ees won’t look at the pho­tos that the police upload.

Clearview also hired Paul D. Clement, a Unit­ed States solic­i­tor gen­er­al under Pres­i­dent George W. Bush, to assuage con­cerns about the app’s legal­i­ty.

In an August memo that Clearview pro­vid­ed to poten­tial cus­tomers, includ­ing the Atlanta Police Depart­ment and the Pinel­las Coun­ty Sheriff’s Office in Flori­da, Mr. Clement said law enforce­ment agen­cies “do not vio­late the fed­er­al Con­sti­tu­tion or rel­e­vant exist­ing state bio­met­ric and pri­va­cy laws when using Clearview for its intend­ed pur­pose.”

Mr. Clement, now a part­ner at Kirk­land & Ellis, wrote that the author­i­ties don’t have to tell defen­dants that they were iden­ti­fied via Clearview, as long as it isn’t the sole basis for get­ting a war­rant to arrest them. Mr. Clement did not respond to mul­ti­ple requests for com­ment.

The memo appeared to be effec­tive; the Atlanta police and Pinel­las Coun­ty Sheriff’s Office soon start­ed using Clearview.

Because the police upload pho­tos of peo­ple they’re try­ing to iden­ti­fy, Clearview pos­sess­es a grow­ing data­base of indi­vid­u­als who have attract­ed atten­tion from law enforce­ment. The com­pa­ny also has the abil­i­ty to manip­u­late the results that the police see. After the com­pa­ny real­ized I was ask­ing offi­cers to run my pho­to through the app, my face was flagged by Clearview’s sys­tems and for a while showed no match­es. When asked about this, Mr. Ton-That laughed and called it a “soft­ware bug.”

“It’s creepy what they’re doing, but there will be many more of these com­pa­nies. There is no monop­oly on math,” said Al Gidari, a pri­va­cy pro­fes­sor at Stan­ford Law School. “Absent a very strong fed­er­al pri­va­cy law, we’re all screwed.”

Mr. Ton-That said his com­pa­ny used only pub­licly avail­able images. If you change a pri­va­cy set­ting in Face­book so that search engines can’t link to your pro­file, your Face­book pho­tos won’t be includ­ed in the data­base, he said.

But if your pro­file has already been scraped, it is too late. The com­pa­ny keeps all the images it has scraped even if they are lat­er delet­ed or tak­en down, though Mr. Ton-That said the com­pa­ny was work­ing on a tool that would let peo­ple request that images be removed if they had been tak­en down from the web­site of ori­gin.

Woodrow Hart­zog, a pro­fes­sor of law and com­put­er sci­ence at North­east­ern Uni­ver­si­ty in Boston, sees Clearview as the lat­est proof that facial recog­ni­tion should be banned in the Unit­ed States.

“We’ve relied on indus­try efforts to self-police and not embrace such a risky tech­nol­o­gy, but now those dams are break­ing because there is so much mon­ey on the table,” Mr. Hart­zog said. “I don’t see a future where we har­ness the ben­e­fits of face recog­ni­tion tech­nol­o­gy with­out the crip­pling abuse of the sur­veil­lance that comes with it. The only way to stop it is to ban it.”

Where Every­body Knows Your Name

Dur­ing a recent inter­view at Clearview’s offices in a WeWork loca­tion in Manhattan’s Chelsea neigh­bor­hood, Mr. Ton-That demon­strat­ed the app on him­self. He took a self­ie and uploaded it. The app pulled up 23 pho­tos of him. In one, he is shirt­less and light­ing a cig­a­rette while cov­ered in what looks like blood.

Mr. Ton-That then took my pho­to with the app. The “soft­ware bug” had been fixed, and now my pho­to returned numer­ous results, dat­ing back a decade, includ­ing pho­tos of myself that I had nev­er seen before. When I used my hand to cov­er my nose and the bot­tom of my face, the app still returned sev­en cor­rect match­es for me.

Police offi­cers and Clearview’s investors pre­dict that its app will even­tu­al­ly be avail­able to the pub­lic.

Mr. Ton-That said he was reluc­tant. “There’s always going to be a com­mu­ni­ty of bad peo­ple who will mis­use it,” he said.

Even if Clearview doesn’t make its app pub­licly avail­able, a copy­cat com­pa­ny might, now that the taboo is bro­ken. Search­ing some­one by face could become as easy as Googling a name. Strangers would be able to lis­ten in on sen­si­tive con­ver­sa­tions, take pho­tos of the par­tic­i­pants and know per­son­al secrets. Some­one walk­ing down the street would be imme­di­ate­ly iden­ti­fi­able — and his or her home address would be only a few clicks away. It would her­ald the end of pub­lic anonymi­ty.

Asked about the impli­ca­tions of bring­ing such a pow­er into the world, Mr. Ton-That seemed tak­en aback.

“I have to think about that,” he said. “Our belief is that this is the best use of the tech­nol­o­gy.”

———–

“The Secre­tive Com­pa­ny That Might End Pri­va­cy as We Know It” by Kash­mir Hill; The New York Times; 01/18/2020

“Even if Clearview doesn’t make its app pub­licly avail­able, a copy­cat com­pa­ny might, now that the taboo is bro­ken. Search­ing some­one by face could become as easy as Googling a name. Strangers would be able to lis­ten in on sen­si­tive con­ver­sa­tions, take pho­tos of the par­tic­i­pants and know per­son­al secrets. Some­one walk­ing down the street would be imme­di­ate­ly iden­ti­fi­able — and his or her home address would be only a few clicks away. It would her­ald the end of pub­lic anonymi­ty.”

An end to pri­va­cy as we know it. Every­one will be able to just look at some­one and imme­di­ate­ly access a data­base of per­son­al infor­ma­tion about them. That’s the dark path Clearview’s tech­nol­o­gy is send­ing us down:

...
But with­out pub­lic scruti­ny, more than 600 law enforce­ment agen­cies have start­ed using Clearview in the past year, accord­ing to the com­pa­ny, which declined to pro­vide a list. The com­put­er code under­ly­ing its app, ana­lyzed by The New York Times, includes pro­gram­ming lan­guage to pair it with aug­ment­ed-real­i­ty glass­es; users would poten­tial­ly be able to iden­ti­fy every per­son they saw. The tool could iden­ti­fy activists at a protest or an attrac­tive stranger on the sub­way, reveal­ing not just their names but where they lived, what they did and whom they knew.

And it’s not just law enforce­ment: Clearview has also licensed the app to at least a hand­ful of com­pa­nies for secu­ri­ty pur­pos­es.

“The weaponiza­tion pos­si­bil­i­ties of this are end­less,” said Eric Gold­man, co-direc­tor of the High Tech Law Insti­tute at San­ta Clara Uni­ver­si­ty. “Imag­ine a rogue law enforce­ment offi­cer who wants to stalk poten­tial roman­tic part­ners, or a for­eign gov­ern­ment using this to dig up secrets about peo­ple to black­mail them or throw them in jail.”
...

And both police offi­cer and Clearview’s own investors pre­dict that its app will even­tu­al­ly be avail­able to the pub­lic. And yet if you ask investor David Scal­zo about the pri­va­cy con­cerns, he appears to take the stance that it’s sim­ply impos­si­ble to ban this use of this tech­nol­o­gy whether or leads to a dystopi­an future or not:

...
Anoth­er ear­ly investor is a small firm called Kire­na­ga Part­ners. Its founder, David Scal­zo, dis­missed con­cerns about Clearview mak­ing the inter­net search­able by face, say­ing it’s a valu­able crime-solv­ing tool.

“I’ve come to the con­clu­sion that because infor­ma­tion con­stant­ly increas­es, there’s nev­er going to be pri­va­cy,” Mr. Scal­zo said. “Laws have to deter­mine what’s legal, but you can’t ban tech­nol­o­gy. Sure, that might lead to a dystopi­an future or some­thing, but you can’t ban it.”

...

“It’s creepy what they’re doing, but there will be many more of these com­pa­nies. There is no monop­oly on math,” said Al Gidari, a pri­va­cy pro­fes­sor at Stan­ford Law School. “Absent a very strong fed­er­al pri­va­cy law, we’re all screwed.”

...

Woodrow Hart­zog, a pro­fes­sor of law and com­put­er sci­ence at North­east­ern Uni­ver­si­ty in Boston, sees Clearview as the lat­est proof that facial recog­ni­tion should be banned in the Unit­ed States.

“We’ve relied on indus­try efforts to self-police and not embrace such a risky tech­nol­o­gy, but now those dams are break­ing because there is so much mon­ey on the table,” Mr. Hart­zog said. “I don’t see a future where we har­ness the ben­e­fits of face recog­ni­tion tech­nol­o­gy with­out the crip­pling abuse of the sur­veil­lance that comes with it. The only way to stop it is to ban it.”

...

Police offi­cers and Clearview’s investors pre­dict that its app will even­tu­al­ly be avail­able to the pub­lic.

Mr. Ton-That said he was reluc­tant. “There’s always going to be a com­mu­ni­ty of bad peo­ple who will mis­use it,” he said.

...

Asked about the impli­ca­tions of bring­ing such a pow­er into the world, Mr. Ton-That seemed tak­en aback.

“I have to think about that,” he said. “Our belief is that this is the best use of the tech­nol­o­gy.”
...

But while Clearview’s investors appear to have no prob­lem at all with blaz­ing the trail of this dystopi­an post-pri­va­cy future, the com­pa­ny itself has tak­en pains to get as lit­tle expo­sure as pos­si­ble. It even freaked out when a jour­nal­ist’s pho­to was run through the sys­tem:

...
Clearview has shroud­ed itself in secre­cy, avoid­ing debate about its bound­ary-push­ing tech­nol­o­gy. When I began look­ing into the com­pa­ny in Novem­ber, its web­site was a bare page show­ing a nonex­is­tent Man­hat­tan address as its place of busi­ness. The company’s one employ­ee list­ed on LinkedIn, a sales man­ag­er named “John Good,” turned out to be Mr. Ton-That, using a fake name. For a month, peo­ple affil­i­at­ed with the com­pa­ny would not return my emails or phone calls.

While the com­pa­ny was dodg­ing me, it was also mon­i­tor­ing me. At my request, a num­ber of police offi­cers had run my pho­to through the Clearview app. They soon received phone calls from com­pa­ny rep­re­sen­ta­tives ask­ing if they were talk­ing to the media — a sign that Clearview has the abil­i­ty and, in this case, the appetite to mon­i­tor whom law enforce­ment is search­ing for.

...

The com­pa­ny even­tu­al­ly start­ed answer­ing my ques­tions, say­ing that its ear­li­er silence was typ­i­cal of an ear­ly-stage start-up in stealth mode. Mr. Ton-That acknowl­edged design­ing a pro­to­type for use with aug­ment­ed-real­i­ty glass­es but said the com­pa­ny had no plans to release it. And he said my pho­to had rung alarm bells because the app “flags pos­si­ble anom­alous search behav­ior” in order to pre­vent users from con­duct­ing what it deemed “inap­pro­pri­ate search­es.”

...

Because the police upload pho­tos of peo­ple they’re try­ing to iden­ti­fy, Clearview pos­sess­es a grow­ing data­base of indi­vid­u­als who have attract­ed atten­tion from law enforce­ment. The com­pa­ny also has the abil­i­ty to manip­u­late the results that the police see. After the com­pa­ny real­ized I was ask­ing offi­cers to run my pho­to through the app, my face was flagged by Clearview’s sys­tems and for a while showed no match­es. When asked about this, Mr. Ton-That laughed and called it a “soft­ware bug.”
...

Beyond that, the tech­nol­o­gy has­n’t even been val­i­dat­ed. Accord­ing to Ton-That, it works about 75 per­cent of the time, which sounds pret­ty good until you real­ize you’re talk­ing about mis­match­es that could lead to the wrong per­son being arrest­ed and charged with a crime:

...
Facial recog­ni­tion tech­nol­o­gy has always been con­tro­ver­sial. It makes peo­ple ner­vous about Big Broth­er. It has a ten­den­cy to deliv­er false match­es for cer­tain groups, like peo­ple of col­or. And some facial recog­ni­tion prod­ucts used by the police — includ­ing Clearview’s — haven’t been vet­ted by inde­pen­dent experts.

Clearview’s app car­ries extra risks because law enforce­ment agen­cies are upload­ing sen­si­tive pho­tos to the servers of a com­pa­ny whose abil­i­ty to pro­tect its data is untest­ed.

...

Anoth­er engi­neer was hired to per­fect a facial recog­ni­tion algo­rithm that was derived from aca­d­e­m­ic papers. The result: a sys­tem that uses what Mr. Ton-That described as a “state-of-the-art neur­al net” to con­vert all the images into math­e­mat­i­cal for­mu­las, or vec­tors, based on facial geom­e­try — like how far apart a person’s eyes are. Clearview cre­at­ed a vast direc­to­ry that clus­tered all the pho­tos with sim­i­lar vec­tors into “neigh­bor­hoods.” When a user uploads a pho­to of a face into Clearview’s sys­tem, it con­verts the face into a vec­tor and then shows all the scraped pho­tos stored in that vector’s neigh­bor­hood — along with the links to the sites from which those images came.

...

Mr. Ton-That said the tool does not always work. Most of the pho­tos in Clearview’s data­base are tak­en at eye lev­el. Much of the mate­r­i­al that the police upload is from sur­veil­lance cam­eras mount­ed on ceil­ings or high on walls.

“They put sur­veil­lance cam­eras too high,” Mr. Ton-That lament­ed. “The angle is wrong for good face recog­ni­tion.”

Despite that, the com­pa­ny said, its tool finds match­es up to 75 per­cent of the time. But it is unclear how often the tool deliv­ers false match­es, because it has not been test­ed by an inde­pen­dent par­ty such as the Nation­al Insti­tute of Stan­dards and Tech­nol­o­gy, a fed­er­al agency that rates the per­for­mance of facial recog­ni­tion algo­rithms.

“We have no data to sug­gest this tool is accu­rate,” said Clare Garvie, a researcher at George­town University’s Cen­ter on Pri­va­cy and Tech­nol­o­gy, who has stud­ied the government’s use of facial recog­ni­tion. “The larg­er the data­base, the larg­er the risk of misiden­ti­fi­ca­tion because of the dop­pel­gänger effect. They’re talk­ing about a mas­sive data­base of ran­dom peo­ple they’ve found on the inter­net.”
...

But while the pos­si­ble mis­us­es of unproven tech­nol­o­gy by law enforce­ment is obvi­ous­ly a prob­lem, it’s the fact that the com­pa­ny appears to run by par­ti­san Repub­li­cans that points towards one of the biggest poten­tial sources of abuse. It’s a polit­i­cal oppo­si­tion research dream tool and the firm is using Repub­li­cans find clients:

...
In addi­tion to Mr. Ton-That, Clearview was found­ed by Richard Schwartz — who was an aide to Rudolph W. Giu­liani when he was may­or of New York — and backed finan­cial­ly by Peter Thiel, a ven­ture cap­i­tal­ist behind Face­book and Palan­tir.

Anoth­er ear­ly investor is a small firm called Kire­na­ga Part­ners. Its founder, David Scal­zo, dis­missed con­cerns about Clearview mak­ing the inter­net search­able by face, say­ing it’s a valu­able crime-solv­ing tool.

...

Mr. Schwartz and Mr. Ton-That met in 2016 at a book event at the Man­hat­tan Insti­tute, a con­ser­v­a­tive think tank. Mr. Schwartz, now 61, had amassed an impres­sive Rolodex work­ing for Mr. Giu­liani in the 1990s and serv­ing as the edi­to­r­i­al page edi­tor of The New York Dai­ly News in the ear­ly 2000s. The two soon decid­ed to go into the facial recog­ni­tion busi­ness togeth­er: Mr. Ton-That would build the app, and Mr. Schwartz would use his con­tacts to drum up com­mer­cial inter­est.

...

Mr. Schwartz paid for serv­er costs and basic expens­es, but the oper­a­tion was bare bones; every­one worked from home. “I was liv­ing on cred­it card debt,” Mr. Ton-That said. “Plus, I was a Bit­coin believ­er, so I had some of those.”

...

Clearview deployed cur­rent and for­mer Repub­li­can offi­cials to approach police forces, offer­ing free tri­als and annu­al licens­es for as lit­tle as $2,000. Mr. Schwartz tapped his polit­i­cal con­nec­tions to help make gov­ern­ment offi­cials aware of the tool, accord­ing to Mr. Ton-That. (“I’m thrilled to have the oppor­tu­ni­ty to help Hoan build Clearview into a mis­sion-dri­ven orga­ni­za­tion that’s help­ing law enforce­ment pro­tect chil­dren and enhance the safe­ty of com­mu­ni­ties across the coun­try,” Mr. Schwartz said through a spokes­woman.)

The company’s main con­tact for cus­tomers was Jes­si­ca Medeiros Gar­ri­son, who man­aged Luther Strange’s Repub­li­can cam­paign for Alaba­ma attor­ney gen­er­al. Bran­don Fricke, an N.F.L. agent engaged to the Fox Nation host Tomi Lahren, said in a finan­cial dis­clo­sure report dur­ing a con­gres­sion­al cam­paign in Cal­i­for­nia that he was a “growth con­sul­tant” for the com­pa­ny. (Clearview said that it was a brief, unpaid role, and that the com­pa­ny had enlist­ed Democ­rats to help mar­ket its prod­uct as well.)
...

And “extreme oppo­si­tion research” is one of the ser­vices Clearview offered one of its first poten­tial clients. That client hap­pened to be Paul Nehlen, the GOP ris­ing-star who saw his polit­i­cal future implode after it became clear he was an open neo-Nazi. That’s the per­son Clearview offered ser­vices to right after the com­pa­ny fin­ished its ini­tial prod­uct in late 2017 and those ser­vices hap­pened to be “extreme oppo­si­tion research”. It tells us A LOT about the real intent of the fig­ures behind this com­pa­ny. It’s not just for law enforce­ment. It’s also a reminder that the com­pa­ny’s will­ing­ness to manip­u­late the search results could be very hand for right-wing politi­cians who would pre­fer embar­rass­ing pics not be read­i­ly avail­able for oppo­nents to find:

...
By the end of 2017, the com­pa­ny had a for­mi­da­ble facial recog­ni­tion tool, which it called Smartcheckr. But Mr. Schwartz and Mr. Ton-That weren’t sure whom they were going to sell it to.

Maybe it could be used to vet babysit­ters or as an add-on fea­ture for sur­veil­lance cam­eras. What about a tool for secu­ri­ty guards in the lob­bies of build­ings or to help hotels greet guests by name? “We thought of every idea,” Mr. Ton-That said.

One of the odd­er pitch­es, in late 2017, was to Paul Nehlen — an anti-Semi­te and self-described “pro-white” Repub­li­can run­ning for Con­gress in Wis­con­sin — to use “uncon­ven­tion­al data­bas­es” for “extreme oppo­si­tion research,” accord­ing to a doc­u­ment pro­vid­ed to Mr. Nehlen and lat­er post­ed online. Mr. Ton-That said the com­pa­ny nev­er actu­al­ly offered such ser­vices.
...

And then, short­ly after mak­ing that offer to Nehlen, Clearview gets its first out­side invest­ment, includ­ing $200,000 from Peter Thiel that was lat­er con­vert­ed to equi­ty. So in addi­tion to co-found­ing Palan­tir and sit­ting on the board of Face­book, Thiel owns an undis­closed amount of this com­pa­ny too. And Ton-That claims Face­book is aware that Clearview’s data­base is heav­i­ly pop­u­lat­ed with data scraped from Face­book:

...
The com­pa­ny soon changed its name to Clearview AI and began mar­ket­ing to law enforce­ment. That was when the com­pa­ny got its first round of fund­ing from out­side investors: Mr. Thiel and Kire­na­ga Part­ners. Among oth­er things, Mr. Thiel was famous for secret­ly financ­ing Hulk Hogan’s law­suit that bank­rupt­ed the pop­u­lar web­site Gawk­er. Both Mr. Thiel and Mr. Ton-That had been the sub­ject of neg­a­tive arti­cles by Gawk­er.

“In 2017, Peter gave a tal­ent­ed young founder $200,000, which two years lat­er con­vert­ed to equi­ty in Clearview AI,” said Jere­mi­ah Hall, Mr. Thiel’s spokesman. “That was Peter’s only con­tri­bu­tion; he is not involved in the com­pa­ny.”

Even after a sec­ond fund­ing round in 2019, Clearview remains tiny, hav­ing raised $7 mil­lion from investors, accord­ing to Pitch­book, a web­site that tracks invest­ments in start-ups. The com­pa­ny declined to con­firm the amount.

...

One rea­son that Clearview is catch­ing on is that its ser­vice is unique. That’s because Face­book and oth­er social media sites pro­hib­it peo­ple from scrap­ing users’ images — Clearview is vio­lat­ing the sites’ terms of ser­vice.

“A lot of peo­ple are doing it,” Mr. Ton-That shrugged. “Face­book knows.”

Jay Nan­car­row, a Face­book spokesman, said the com­pa­ny was review­ing the sit­u­a­tion with Clearview and “will take appro­pri­ate action if we find they are vio­lat­ing our rules.”

Mr. Thiel, the Clearview investor, sits on Facebook’s board. Mr. Nan­car­row declined to com­ment on Mr. Thiel’s per­son­al invest­ments.

...

Mr. Ton-That said his com­pa­ny used only pub­licly avail­able images. If you change a pri­va­cy set­ting in Face­book so that search engines can’t link to your pro­file, your Face­book pho­tos won’t be includ­ed in the data­base, he said.

But if your pro­file has already been scraped, it is too late. The com­pa­ny keeps all the images it has scraped even if they are lat­er delet­ed or tak­en down, though Mr. Ton-That said the com­pa­ny was work­ing on a tool that would let peo­ple request that images be removed if they had been tak­en down from the web­site of ori­gin.
...

Beyond that, Clearview hired high-pro­file Repub­li­can lawyer Paul D. Clement to assure clients that the ser­vices are legal. Giv­en who Clement is in the legal world that’s a major legal endorse­ment:

...
Clearview also hired Paul D. Clement, a Unit­ed States solic­i­tor gen­er­al under Pres­i­dent George W. Bush, to assuage con­cerns about the app’s legal­i­ty.

In an August memo that Clearview pro­vid­ed to poten­tial cus­tomers, includ­ing the Atlanta Police Depart­ment and the Pinel­las Coun­ty Sheriff’s Office in Flori­da, Mr. Clement said law enforce­ment agen­cies “do not vio­late the fed­er­al Con­sti­tu­tion or rel­e­vant exist­ing state bio­met­ric and pri­va­cy laws when using Clearview for its intend­ed pur­pose.”

Mr. Clement, now a part­ner at Kirk­land & Ellis, wrote that the author­i­ties don’t have to tell defen­dants that they were iden­ti­fied via Clearview, as long as it isn’t the sole basis for get­ting a war­rant to arrest them. Mr. Clement did not respond to mul­ti­ple requests for com­ment.

The memo appeared to be effec­tive; the Atlanta police and Pinel­las Coun­ty Sheriff’s Office soon start­ed using Clearview.
...

Oh, and it turns out the FBI and DHS are also try­ing out Clearview’s ser­vices, along with Cana­di­an law enforce­ment agen­cies:

...
Fed­er­al law enforce­ment, includ­ing the F.B.I. and the Depart­ment of Home­land Secu­ri­ty, are try­ing it, as are Cana­di­an law enforce­ment author­i­ties, accord­ing to the com­pa­ny and gov­ern­ment offi­cials.
...

It does­n’t sound like fed­er­al agen­cies have a prob­lem with using a data­base of images that was improp­er­ly scraped off of major social media sites. That’s appar­ent­ly legal and fine. And that’s why Clearview appears to be on track to becom­ing the ‘Palan­tir’ of facial recog­ni­tion com­pa­nies: a high­ly secre­tive com­pa­ny owned by polit­i­cal con­nect­ed shady fig­ures that some­how man­ages to get mas­sive num­bers of gov­ern­ment clients by offer­ing ser­vices that have obvi­ous intel­li­gence appli­ca­tions. And it’s co-owned by Peter Thiel, fur­ther solid­i­fy­ing Thiel’s posi­tion as the US’s pri­vate intel­li­gence oli­garch. It’s quite a posi­tion for an open fas­cist like Thiel.

So when ser­vices like this end pri­va­cy as we know it ends soon­er than you expect, don’t for­get that this was brought to you by Repub­li­cans who did­n’t want you to know about those ser­vices in the first place.

Oh look at that, Face­book just hired a new head of video strat­e­gy per­son to head up the video divi­sion for the “Face­book News” fea­ture that its cre­at­ing for 2020. Guess who: Jen­nifer William, an 18-year vet­er­an of Fox News. Sur­prise!

And Williams isn’t just a Fox News vet­er­an. She’s was a long-time senior pro­duc­er of Fox & Friends (from 1997–2009), one of the chan­nel’s most egre­gious out­lets of dis­in­for­ma­tion. Fox & Friends is bad even by Fox News stan­dards. That’s who is head­ing up the video sec­tion of Face­book’s new News sec­tion:

Media Mat­ters

After enabling right-wing pro­pa­gan­da, Face­book hires a Fox News vet­er­an in a key news role

Social media giant again moves to bol­ster the right as 2020 elec­tions loom

Writ­ten by Matt Gertz
Pub­lished 01/28/20 12:01 PM EST

In 2007, after Fox & Friends pro­mot­ed a quick­ly debunked report that then-Sen. Barack Oba­ma had gone to school at an extrem­ist Islam­ic madras­sa as a child, a top Fox News exec­u­tive issued a tru­ly star­tling inter­nal memo to the network’s news­room. “For the record,” then-Vice Pres­i­dent of News John Moody wrote, “see­ing an item on a web­site does not mean it is right. Nor does it mean it is ready for air on FNC.”

Thir­teen years lat­er, Face­book has report­ed­ly named Jen­nifer Williams, who was a Fox & Friends senior pro­duc­er at the time that memo was sent, to head video strat­e­gy for the social media giant’s forth­com­ing Face­book News, NBC News report­ed Tues­day. Face­book News will serve its bil­lions of users with a ded­i­cat­ed tab includ­ing news con­tent curat­ed by a team of jour­nal­ists from a list of pub­lish­ers cho­sen by the com­pa­ny. As Face­book exec­u­tives plan a shift in the way the nation con­sumes news that will almost cer­tain­ly impact the 2020 pres­i­den­tial elec­tions, they are staffing up with an 18-year vet­er­an of the right-wing cable net­work that effec­tive­ly serves as Pres­i­dent Don­ald Trump’s per­son­al mouth­piece.

Facebook’s mas­sive audi­ence and immense pow­er have placed the com­pa­ny at the cen­ter of a polit­i­cal mael­strom over the last half decade. The 2016 deci­sion by its CEO, Mark Zucker­berg, to bow to a bogus right-wing pres­sure cam­paign and elim­i­nate the human cura­tors who man­aged its Trend­ing Top­ics sec­tion helped turn the plat­form into a tox­ic fake news ecosys­tem rid­dled with for­eign inter­fer­ence in the days lead­ing up to Trump’s elec­tion.

That campaign’s suc­cess only embold­ened the right, which has con­tin­ued to offer flawed attacks on Facebook’s pur­port­ed anti-con­ser­v­a­tive bias even as the plat­form has “become a hub for some of the most insid­i­ous con­ser­v­a­tive false talk­ing points about immi­gra­tion, abor­tion, cli­mate change, and trans rights,” as my col­league Park­er Mol­loy has not­ed. (Media Mat­ters has pub­lished mul­ti­ple stud­ies demon­strat­ing that con­ser­v­a­tives are not cen­sored on Face­book.) And while fac­ing a left that threat­ens to aggres­sive­ly reg­u­late or even break up the com­pa­ny — and the prospect of high­er tax­es for its top exec­u­tives under a Demo­c­ra­t­ic admin­is­tra­tion — Face­book has repeat­ed­ly act­ed to bol­ster Trump and oth­er con­ser­v­a­tives.

The hir­ing of Williams, who spent 12 years at Fox’s insipid morn­ing show and lat­er served as exec­u­tive pro­duc­er for Gretchen Carl­son and Bush admin­is­tra­tion press sec­re­tary Dana Peri­no, is sim­ply the lat­est exam­ple of Face­book sid­ing with the right. While Zucker­berg has been hold­ing cozy din­ners with right-wing media fig­ures and meet­ing with Trump in the White House, his com­pa­ny has stocked its pow­er­ful Wash­ing­ton, D.C., office with Repub­li­can polit­i­cal oper­a­tives, hired a for­mer Repub­li­can sen­a­tor to pro­duce a report on Facebook’s alleged lib­er­al bias, encour­aged its fact-check­ing con­sor­tium to part­ner with Tuck­er Carlson’s tox­ic Dai­ly Caller web­site, com­mis­sioned research on oppo­nents that was steeped in anti-Semit­ic tropes, and altered and selec­tive­ly enforced its ad poli­cies in ways that will like­ly ben­e­fit Trump’s reelec­tion cam­paign. Just last week, Pop­u­lar Infor­ma­tion’s Judd Legum report­ed that Face­book was allow­ing a major Trump super PAC to run ads fea­tur­ing a lie that the plat­for­m’s own fact-check­ing part­ners had debunked.

Face­book News pos­es a new chal­lenge for the com­pa­ny, requir­ing it to make affir­ma­tive jour­nal­is­tic deci­sions rather than allow­ing its algo­rithm to deter­mine what users see. Its exec­u­tives will need to decide if they want to sup­port a fac­tu­al news envi­ron­ment, or cater to a right-wing media that often pro­duces con­spir­a­cy the­o­ries and big­ot­ed con­tent.

Face­book hasn’t rolled out the news tab yet, but ear­ly indi­ca­tions sug­gest the com­pa­ny may plan to take the lat­ter path. Camp­bell Brown, a for­mer NBC News anchor who has been at Face­book since 2017, has been tapped to lead the pro­gram. Before join­ing Face­book, Brown served as edi­tor-in-chief of The 74, an edu­ca­tion pol­i­cy news web­site that was fund­ed by the fam­i­ly foun­da­tion of Bet­sy DeVos, who sub­se­quent­ly joined Trump’s cab­i­net, as Judd Legum has report­ed. And under Brown’s tenure, Face­book News has cre­den­tialed the nox­ious hate site Breitbart.com as one of the sites the tab will pro­mote. Brown’s defense of that deci­sion — that Breitbart.com “meets our integri­ty stan­dards for mis­in­for­ma­tion” — sug­gests that those stan­dards will be absurd­ly low for right-wing media.

...

———-

“After enabling right-wing pro­pa­gan­da, Face­book hires a Fox News vet­er­an in a key news role” by Matt Gertz; Media Mat­ters; 01/28/2020

“Thir­teen years lat­er, Face­book has report­ed­ly named Jen­nifer Williams, who was a Fox & Friends senior pro­duc­er at the time that memo was sent, to head video strat­e­gy for the social media giant’s forth­com­ing Face­book News, NBC News report­ed Tues­day. Face­book News will serve its bil­lions of users with a ded­i­cat­ed tab includ­ing news con­tent curat­ed by a team of jour­nal­ists from a list of pub­lish­ers cho­sen by the com­pa­ny. As Face­book exec­u­tives plan a shift in the way the nation con­sumes news that will almost cer­tain­ly impact the 2020 pres­i­den­tial elec­tions, they are staffing up with an 18-year vet­er­an of the right-wing cable net­work that effec­tive­ly serves as Pres­i­dent Don­ald Trump’s per­son­al mouth­piece.”

An 18-year vet­er­an of Fox News. That’s who is going to be ulti­mate­ly curat­ing the ‘news’ videos served up to Face­book read­ers. As the arti­cle notes, the fact that Face­book decid­ed to make a spe­cial ‘news’ sec­tion osten­si­bly man­aged with a jour­nal­is­tic intent behind it, and not just run by an algo­rithm, meant the com­pa­ny was going to have to get in the busi­ness of hav­ing humans make active deci­sions on whether or not news is wor­thy of being includ­ed in the new News sec­tion of the site or if it’s ‘fake news’. So Face­book chose the vet­er­an of the lead­ing pur­vey­or fake news.

But Jen­nifer Williams isn’t the only high­ly ques­tion­able fig­ure who is going to be run­ning Face­book’s new News divi­sion. the com­pa­ny already hired Camp­bell Brown to lead the News divi­sion. And it turns out Brown is close to Bet­sy DeVos, the far right sis­ter of Erik Prince and Trump’s Edu­ca­tion Sec­re­tary. As we should expect, Brown has already decid­ed to cre­den­tial Breibart.com as one of the new sites that Face­book News will pro­mote:

...
Face­book News pos­es a new chal­lenge for the com­pa­ny, requir­ing it to make affir­ma­tive jour­nal­is­tic deci­sions rather than allow­ing its algo­rithm to deter­mine what users see. Its exec­u­tives will need to decide if they want to sup­port a fac­tu­al news envi­ron­ment, or cater to a right-wing media that often pro­duces con­spir­a­cy the­o­ries and big­ot­ed con­tent.

Face­book hasn’t rolled out the news tab yet, but ear­ly indi­ca­tions sug­gest the com­pa­ny may plan to take the lat­ter path. Camp­bell Brown, a for­mer NBC News anchor who has been at Face­book since 2017, has been tapped to lead the pro­gram. Before join­ing Face­book, Brown served as edi­tor-in-chief of The 74, an edu­ca­tion pol­i­cy news web­site that was fund­ed by the fam­i­ly foun­da­tion of Bet­sy DeVos, who sub­se­quent­ly joined Trump’s cab­i­net, as Judd Legum has report­ed. And under Brown’s tenure, Face­book News has cre­den­tialed the nox­ious hate site Breitbart.com as one of the sites the tab will pro­mote. Brown’s defense of that deci­sion — that Breitbart.com “meets our integri­ty stan­dards for mis­in­for­ma­tion” — sug­gests that those stan­dards will be absurd­ly low for right-wing media.
...

And now here’s Judd Legum’s Popular.Info piece with more on Cam­bell Brown and her exten­sive ties to Bet­sy DeVos. As the arti­cle notes, the pub­li­ca­tion Brown co-found­ed, The 74, is large­ly focused on edu­ca­tion news, which got rather awk­ward after Bet­sy DeVos became Trump’s edu­ca­tion sec­re­tary. DeVos calls Brown a “friend” and The 74 was start­ed, in part, with a $200,000 grant from Bet­sy DeVos’s fam­i­ly foun­da­tion. Most of the arti­cles in The 74 cov­er­ing DeVos have been large­ly lauda­to­ry.
Brown is also a mem­ber of the board of The Amer­i­can Fed­er­a­tion for Chil­dren (AFC), a right-wing non-prof­it start­ed and chaired by DeVos that spends heav­i­ly on get­ting Repub­li­cans elect­ed at the state lev­el. The 74 and the AFC co-spon­sored a Repub­li­can pres­i­den­tial forum in Iowa in 2015.

It’s also worth recall­ing the recent sto­ry describ­ing how the DeVos’s and oth­er far right oli­garchs asso­ci­at­ed with the theo­crat­ic Coun­cil for Nation­al Pol­i­cy (CNP) have been qui­et­ly financ­ing the pur­chase of local and region­al radio sta­tions to ensure the explo­sive growth of region­al right-wing talk radio. It’s a reminder that the dam­age Bet­sy DeVos is doing to the intel­lec­tu­al sta­tus of Amer­i­ca isn’t lim­it­ed to the dam­age she’s doing to Amer­i­can edu­ca­tion.

As anoth­er sign of Brown’s edi­to­r­i­al lean­ings, while edi­tor-in-chief of The 74, the pub­li­ca­tion fea­tured at least 11 pieces from Eric Owens, a Dai­ly Caller edi­tor with a long his­to­ry of mak­ing trans­pho­bic attacks on stu­dents and teach­ers. The 74 also appears to real­ly hate Eliz­a­beth War­ren. Inter­est­ing­ly, Mark Zucker­berg’s foun­da­tion, the Chan Zucker­berg Ini­tia­tive, donat­ed $600,000 to The 74, describ­ing it as “a non-prof­it, non­par­ti­san news site cov­er­ing edu­ca­tion in Amer­i­ca.” Zucker­berg has pre­vi­ous­ly expressed his extreme dis­like of War­ren’s pres­i­den­tial ambi­tions, describ­ing her as an “exis­ten­tial” threat to the com­pa­ny.

So that’s who Jen­nifer Williams is going to be report­ing to in her new role as the head of Video at Face­book News: Cam­bell Brown, the right-wing friend of Bet­sy DeVos:

Popular.info

Face­book’s top news exec­u­tive has her own media out­let — and it’s been sav­aging Eliz­a­beth War­ren

Judd Legum
Nov 11, 2019

For­mer NBC News anchor Camp­bell Brown is a top Face­book exec­u­tive who was hired in Jan­u­ary 2017 to lead the com­pa­ny’s “news part­ner­ships team.” That means Brown is in charge of “Face­book News,” the com­pa­ny’s high-pro­file new effort to fea­ture “qual­i­ty news” in a ded­i­cat­ed tab. She is also a co-founder and direc­tor of her own media out­let that, in recent weeks, has harsh­ly attacked one of the lead­ing Demo­c­ra­t­ic can­di­dates for pres­i­dent, Eliz­a­beth War­ren.

In 2015, Brown co-found­ed The 74, which focus­es on the pub­lic edu­ca­tion sys­tem, and served as edi­tor-in-chief. Even after join­ing Face­book in 2017, Brown has main­tained an active role in The 74, where she is a mem­ber of the board of direc­tors. Accord­ing to doc­u­ments filed with the IRS in 2017, Brown ded­i­cat­ed five hours per week — the equiv­a­lent of a month-and-a-half of full-time work — work­ing for The 74.

[see image of IRS for­mer indi­cat­ing Brown was work­ing for The 74 for 5 hours a week]

That’s the same amount of time Brown spent on The 74 pri­or to join­ing Face­book. (2017 is the most recent year that this infor­ma­tion is pub­licly avail­able.)

Begin­ning this fall, The 74 has harsh­ly crit­i­cized pres­i­den­tial can­di­date Eliz­a­beth War­ren. On Octo­ber 23, The 74 pub­lished an arti­cle with this head­line: “Eliz­a­beth Warren’s Edu­ca­tion Plan Is Exact­ly What We Need — If Our Goal Is to Make the Achieve­ment Gap Per­ma­nent.” The piece described War­ren’s detailed edu­ca­tion plan as “a cut-and-paste gen­u­flect to the pub­lic rela­tions depart­ments of America’s nation­al teach­ers unions.” It goes on to claim that War­ren is not a “straight shoot­er” and lacks a “moral cen­ter.” The piece even­tu­al­ly dis­pens­es with edu­ca­tion pol­i­cy alto­geth­er and launch­es into a dia­tribe of attacks on War­ren:

She’s a mil­lion­aire who raves about social­ism. She was Repub­li­can before she was a Demo­c­rat. She was for school choice before she was against it. She was for char­ter schools before she was against them. She was for stan­dard­ized test­ing before she was against it.

She was Native Amer­i­can before she wasn’t.

This piece is not an aber­ra­tion. An Octo­ber 10 piece described War­ren as “the sec­ond com­ing of Karl Marx.”

An Octo­ber 24 col­umn accus­es War­ren of stand­ing “against an insti­tu­tion designed to cre­ate oppor­tu­ni­ty for our nation’s chil­dren.” War­ren, accord­ing to the arti­cle, wants to “over­ride… the clear pref­er­ences of the black and brown vot­ers whom pro­gres­sives claim to fight for.”

An Octo­ber 28 piece describes War­ren as “anoth­er tired politi­cian sign­ing up to pledge undy­ing loy­al­ty to a sys­tem that is so clear­ly fail­ing too many of our chil­dren.” The col­umn says War­ren backs “the regres­sive sta­tus quo that leads our chil­dren into the school-to-prison pipeline.”

Brown fea­tures her affil­i­a­tion with The 74 on both her Face­book page and her Twit­ter pro­file.

Face­book did not answer a detailed set of ques­tions about Brown’s cur­rent duties at The 74 and whether there was a con­flict with her work at Face­book. But the com­pa­ny sent Pop­u­lar Infor­ma­tion the fol­low­ing state­ment: “The 74 is not part of Face­book News. Campbell’s work with The 74 is well-known and she’s been trans­par­ent about her role with the non­prof­it for many years.“

Camp­bell Brown’s friend Bet­sy DeVos

Both Brown and The 74 are tight­ly linked to Bet­sy DeVos, Trump’s Sec­re­tary of Edu­ca­tion. DeVos, who Brown calls a “friend,” pro­vid­ed a two-year grant through her fam­i­ly foun­da­tion to help launch The 74. (The 74 has not dis­closed the amount of DeVos’ con­tri­bu­tion.) Brown also served on the board of The Amer­i­can Fed­er­a­tion for Chil­dren (AFC), a non-prof­it that DeVos found­ed and chaired.

The AFC is a right-wing orga­ni­za­tion that spends heav­i­ly to sup­port Repub­li­cans at the state lev­el. It spent mil­lions, for exam­ple, to sup­port for­mer Wis­con­sin Gov­er­nor Scott Walk­er ® and his allies. The 74 and the AFC co-spon­sored a Repub­li­can pres­i­den­tial forum in Iowa in 2015.

In a a col­umn pub­lished after DeVos’ nom­i­na­tion was announced, Brown lav­ished praise on DeVos. Brown called DeVos “tena­cious in defend­ing the best inter­ests of chil­dren rather than inter­est groups and their polit­i­cal patrons.” She described DeVos as “a born deci­sion-mak­er, thick-skinned, nev­er long dis­cour­aged by set­backs and imper­vi­ous to hos­tile crit­i­cism.”

After DeVos was nom­i­nat­ed by Trump, The 74 began includ­ing a dis­claimer on arti­cles about DeVos, not­ing her role in fund­ing the site. The dis­claimer also said that Brown did not edit sto­ries involv­ing DeVos. That dis­claimer, how­ev­er, last appeared in 2017.

The 74’s cov­er­age of DeVos has been occa­sion­al­ly crit­i­cal, but most­ly lauda­to­ry. Head­lines about DeVos’ tenure as Edu­ca­tion Sec­re­tary on The 74 include:

DeVos Pro­posed $50 Mil­lion for Dis­tricts to Decen­tral­ize Fed­er­al Mon­ey, to Put Schools in the Driver’s Seat. It’s a Smart Idea.

Teach­ers Nation­wide Say Obama’s Dis­ci­pline ‘Reform’ Put Them in Dan­ger. So Why Are the Unions Fight­ing DeVos on Repeal?

Ivan­ka Trump, Bet­sy DeVos Tout STEM Edu­ca­tion to 200 Stu­dents at Air & Space Muse­um

Resis­tance to DeVos Has Obscured the True Record of Michigan’s Strong Char­ter Schools

While DeVos has been exco­ri­at­ed by civ­il rights groups, includ­ing the NAACP, The 74 inter­viewed a civ­il rights leader who praised DeVos. In June, DeVos her­self gave an exclu­sive inter­view to The 74. The inter­view, which did not men­tion DeVos’ con­tro­ver­sial pol­i­cy moves on sex­u­al assault and LGBTQ rights, did not include any dis­clo­sure of DeVos’ pri­or fund­ing for the site.

Brown’s The 74 fea­tured big­ot­ed Dai­ly Caller edi­tor

While Brown served as edi­tor-in-chief of The 74, the site fea­tured at least 11 pieces from Eric Owens, an edi­tor at The Dai­ly Caller. Owens “has a long his­to­ry of pen­ning racial­ly insen­si­tive, sex­ist, and trans­pho­bic attacks on stu­dents and teach­ers.”

Owens, for exam­ple, wrote in The Dai­ly Caller that white priv­i­lege is a “rad­i­cal and bizarre polit­i­cal the­o­ry that white peo­ple enjoy a bunch of won­der­ful priv­i­leges while every­one else suf­fers under the yoke of invis­i­ble oppres­sion.” In anoth­er Dai­ly Caller col­umn, Owens called col­lege stu­dents “del­i­cate, imma­ture wuss­es who become trau­ma­tized, get the vapors and seek pro­fes­sion­al coun­sel­ing any time they face adver­si­ty.”

...

After Brown joined Face­book, The Dai­ly Caller was named an offi­cial Face­book fact-check­ing part­ner, despite The Dai­ly Caller’s his­to­ry of inac­cu­rate report­ing.

Brown thinks Bre­it­bart is a “qual­i­ty” news source

Brown’s role with The 74 rais­es fur­ther ques­tions about the ide­o­log­i­cal under­pin­nings of Face­book’s nascent news tab, which has not been rolled out to all users. Brown’s team elect­ed to include Bre­it­bart — an unre­li­able and nox­ious right-wing site that was lit­er­al­ly caught laun­der­ing white nation­al­ist talk­ing points — among the 200 “qual­i­ty” sources includ­ed in the launch.

On Face­book, Brown defend­ed the deci­sion:

I also believe that in build­ing out a des­ti­na­tion for news on Face­book, we should include con­tent from ide­o­log­i­cal pub­lish­ers on both the left and the right — as long as that con­tent meets our integri­ty stan­dards for mis­in­for­ma­tion. All the con­tent on Face­book News today meets those stan­dards. If a pub­lish­er vio­lates our stan­dards by post­ing mis­in­for­ma­tion or hate speech on our plat­form, they will be removed from Face­book News.

It’s unclear how Bre­it­bart could meet any “integri­ty stan­dard for mis­in­for­ma­tion.” In 2017, for exam­ple, Bre­it­bart “made up a false sto­ry that an immi­grant start­ed dead­ly Sono­ma wild­fires.” The sto­ry, which was not backed by “any evi­dence,” was picked up by oth­er right-wing out­lets like The Drudge Report and InfoWars, and spread quick­ly on Face­book. In 2016, Bre­it­bart dis­patched a reporter to a small Ida­ho town to report on a fake “Mus­lim inva­sion.” It hawks scam cryp­tocur­ren­cies to its read­er­ship.

Bre­it­bart is banned from being cit­ed as a source on Wikipedia. The online ency­clo­pe­dia says Bre­it­bart “should not be used, ever, as a ref­er­ence for facts, due to its unre­li­a­bil­i­ty.” Brown, how­ev­er, believes it is a qual­i­ty news source for Face­book read­ers.

Face­book has refused to release a list of the 200 pub­li­ca­tions approved for inclu­sion in the news tab.

Face­book’s hos­til­i­ty toward War­ren

The 74’s hos­til­i­ty toward War­ren echoes com­ments by Face­book CEO Mark Zucker­berg. In leaked audio of a com­pa­ny meet­ing, Zucker­berg said it would “suck” if War­ren became pres­i­dent because she posed an “exis­ten­tial” threat to the com­pa­ny. Zucker­berg promised to “go to the mat” to fight War­ren’s agen­da. An excerpt:

I mean, if [War­ren] gets elect­ed pres­i­dent, then I would bet that we will have a legal chal­lenge, and I would bet that we will win the legal chal­lenge. And does that still suck for us? Yeah. I mean, I don’t want to have a major law­suit against our own gov­ern­ment. I mean, that’s not the posi­tion that you want to be in when you’re, you know, I mean … it’s like, we care about our coun­try and want to work with our gov­ern­ment and do good things. But look, at the end of the day, if someone’s going to try to threat­en some­thing that exis­ten­tial, you go to the mat and you fight.

After the audio leaked, Zucker­berg did not express regret for trash­ing one of the lead­ing Demo­c­ra­t­ic can­di­dates for pres­i­dent in a com­pa­ny meet­ing. Instead, he linked to a tran­script of the audio from his Face­book page page, call­ing it an “unfil­tered ver­sion of what I’m think­ing and telling employ­ees on a bunch of top­ics.”

Zucker­berg has donat­ed $600,000 to The 74 in 2019 through his foun­da­tion, the Chan Zucker­berg Ini­tia­tive. Zucker­berg’s foun­da­tion describes The 74 as “a non-prof­it, non­par­ti­san news site cov­er­ing edu­ca­tion in Amer­i­ca.”
———–

“Face­book’s top news exec­u­tive has her own media out­let — and it’s been sav­aging Eliz­a­beth War­ren” by Judd Legum; Popular.info; 11/11/2019

“In 2015, Brown co-found­ed The 74, which focus­es on the pub­lic edu­ca­tion sys­tem, and served as edi­tor-in-chief. Even after join­ing Face­book in 2017, Brown has main­tained an active role in The 74, where she is a mem­ber of the board of direc­tors. Accord­ing to doc­u­ments filed with the IRS in 2017, Brown ded­i­cat­ed five hours per week — the equiv­a­lent of a month-and-a-half of full-time work — work­ing for The 74.”

The head of Face­book’s new News fea­ture co-found­ed The 74, a pub­li­ca­tion in 2015 focused on edu­ca­tion and remained on the board of direc­tors even after join­ing Face­book. That would­n’t be a huge deal if The 74 was just a blah non-ide­o­log­i­cal out­let. But it turns out to have been found­ed in part with a grant from Bet­sy DeVos’s fam­i­ly foun­da­tion:

...
Both Brown and The 74 are tight­ly linked to Bet­sy DeVos, Trump’s Sec­re­tary of Edu­ca­tion. DeVos, who Brown calls a “friend,” pro­vid­ed a two-year grant through her fam­i­ly foun­da­tion to help launch The 74. (The 74 has not dis­closed the amount of DeVos’ con­tri­bu­tion.) Brown also served on the board of The Amer­i­can Fed­er­a­tion for Chil­dren (AFC), a non-prof­it that DeVos found­ed and chaired.

The AFC is a right-wing orga­ni­za­tion that spends heav­i­ly to sup­port Repub­li­cans at the state lev­el. It spent mil­lions, for exam­ple, to sup­port for­mer Wis­con­sin Gov­er­nor Scott Walk­er ® and his allies. The 74 and the AFC co-spon­sored a Repub­li­can pres­i­den­tial forum in Iowa in 2015.

In a a col­umn pub­lished after DeVos’ nom­i­na­tion was announced, Brown lav­ished praise on DeVos. Brown called DeVos “tena­cious in defend­ing the best inter­ests of chil­dren rather than inter­est groups and their polit­i­cal patrons.” She described DeVos as “a born deci­sion-mak­er, thick-skinned, nev­er long dis­cour­aged by set­backs and imper­vi­ous to hos­tile crit­i­cism.”

...

While DeVos has been exco­ri­at­ed by civ­il rights groups, includ­ing the NAACP, The 74 inter­viewed a civ­il rights leader who praised DeVos. In June, DeVos her­self gave an exclu­sive inter­view to The 74. The inter­view, which did not men­tion DeVos’ con­tro­ver­sial pol­i­cy moves on sex­u­al assault and LGBTQ rights, did not include any dis­clo­sure of DeVos’ pri­or fund­ing for the site.
...

And the con­tent of The 74 has a clear right-wing ori­en­ta­tion, with arti­cles that describe Eliz­a­beth War­ren as “the sec­ond com­ing of Karl Marx”. And it turns out The 74 received a $600,000 dona­tion from none oth­er than Mark Zucker­berg, who open­ly fears and loathes War­ren:

...
Begin­ning this fall, The 74 has harsh­ly crit­i­cized pres­i­den­tial can­di­date Eliz­a­beth War­ren. On Octo­ber 23, The 74 pub­lished an arti­cle with this head­line: “Eliz­a­beth Warren’s Edu­ca­tion Plan Is Exact­ly What We Need — If Our Goal Is to Make the Achieve­ment Gap Per­ma­nent.” The piece described War­ren’s detailed edu­ca­tion plan as “a cut-and-paste gen­u­flect to the pub­lic rela­tions depart­ments of America’s nation­al teach­ers unions.” It goes on to claim that War­ren is not a “straight shoot­er” and lacks a “moral cen­ter.” The piece even­tu­al­ly dis­pens­es with edu­ca­tion pol­i­cy alto­geth­er and launch­es into a dia­tribe of attacks on War­ren:

She’s a mil­lion­aire who raves about social­ism. She was Repub­li­can before she was a Demo­c­rat. She was for school choice before she was against it. She was for char­ter schools before she was against them. She was for stan­dard­ized test­ing before she was against it.

She was Native Amer­i­can before she wasn’t.

This piece is not an aber­ra­tion. An Octo­ber 10 piece described War­ren as “the sec­ond com­ing of Karl Marx.”

An Octo­ber 24 col­umn accus­es War­ren of stand­ing “against an insti­tu­tion designed to cre­ate oppor­tu­ni­ty for our nation’s chil­dren.” War­ren, accord­ing to the arti­cle, wants to “over­ride… the clear pref­er­ences of the black and brown vot­ers whom pro­gres­sives claim to fight for.”

An Octo­ber 28 piece describes War­ren as “anoth­er tired politi­cian sign­ing up to pledge undy­ing loy­al­ty to a sys­tem that is so clear­ly fail­ing too many of our chil­dren.” The col­umn says War­ren backs “the regres­sive sta­tus quo that leads our chil­dren into the school-to-prison pipeline.”

...

The 74’s hos­til­i­ty toward War­ren echoes com­ments by Face­book CEO Mark Zucker­berg. In leaked audio of a com­pa­ny meet­ing, Zucker­berg said it would “suck” if War­ren became pres­i­dent because she posed an “exis­ten­tial” threat to the com­pa­ny. Zucker­berg promised to “go to the mat” to fight War­ren’s agen­da. An excerpt:

I mean, if [War­ren] gets elect­ed pres­i­dent, then I would bet that we will have a legal chal­lenge, and I would bet that we will win the legal chal­lenge. And does that still suck for us? Yeah. I mean, I don’t want to have a major law­suit against our own gov­ern­ment. I mean, that’s not the posi­tion that you want to be in when you’re, you know, I mean … it’s like, we care about our coun­try and want to work with our gov­ern­ment and do good things. But look, at the end of the day, if someone’s going to try to threat­en some­thing that exis­ten­tial, you go to the mat and you fight.

After the audio leaked, Zucker­berg did not express regret for trash­ing one of the lead­ing Demo­c­ra­t­ic can­di­dates for pres­i­dent in a com­pa­ny meet­ing. Instead, he linked to a tran­script of the audio from his Face­book page page, call­ing it an “unfil­tered ver­sion of what I’m think­ing and telling employ­ees on a bunch of top­ics.”

Zucker­berg has donat­ed $600,000 to The 74 in 2019 through his foun­da­tion, the Chan Zucker­berg Ini­tia­tive. Zucker­berg’s foun­da­tion describes The 74 as “a non-prof­it, non­par­ti­san news site cov­er­ing edu­ca­tion in Amer­i­ca.”
...

Then there’s the fact that The 74 fea­tures writ­ers like Eric Owens, an edi­tor at The Dai­ly Caller. After Brown was hired by Face­book to head up its news divi­sion, The Dai­ly Caller was an an offi­cial fact-check­ing part­ner at Face­book:

...
While Brown served as edi­tor-in-chief of The 74, the site fea­tured at least 11 pieces from Eric Owens, an edi­tor at The Dai­ly Caller. Owens “has a long his­to­ry of pen­ning racial­ly insen­si­tive, sex­ist, and trans­pho­bic attacks on stu­dents and teach­ers.”

Owens, for exam­ple, wrote in The Dai­ly Caller that white priv­i­lege is a “rad­i­cal and bizarre polit­i­cal the­o­ry that white peo­ple enjoy a bunch of won­der­ful priv­i­leges while every­one else suf­fers under the yoke of invis­i­ble oppres­sion.” In anoth­er Dai­ly Caller col­umn, Owens called col­lege stu­dents “del­i­cate, imma­ture wuss­es who become trau­ma­tized, get the vapors and seek pro­fes­sion­al coun­sel­ing any time they face adver­si­ty.”

...

After Brown joined Face­book, The Dai­ly Caller was named an offi­cial Face­book fact-check­ing part­ner, despite The Dai­ly Caller’s his­to­ry of inac­cu­rate report­ing.
...

Oh, and Brown’s team at Face­book end­ed up select­ing Bre­it­bart, which is banned as a cita­tion source for Wikipedia, as one of its 200 “qual­i­ty” news sources:

...
Brown’s role with The 74 rais­es fur­ther ques­tions about the ide­o­log­i­cal under­pin­nings of Face­book’s nascent news tab, which has not been rolled out to all users. Brown’s team elect­ed to include Bre­it­bart — an unre­li­able and nox­ious right-wing site that was lit­er­al­ly caught laun­der­ing white nation­al­ist talk­ing points — among the 200 “qual­i­ty” sources includ­ed in the launch.

...

Bre­it­bart is banned from being cit­ed as a source on Wikipedia. The online ency­clo­pe­dia says Bre­it­bart “should not be used, ever, as a ref­er­ence for facts, due to its unre­li­a­bil­i­ty.” Brown, how­ev­er, believes it is a qual­i­ty news source for Face­book read­ers.

Face­book has refused to release a list of the 200 pub­li­ca­tions approved for inclu­sion in the news tab.
...

So as we can see, the head of Face­book’s new News fea­ture that’s going to roll out some time in 2020 is a close friend of Bet­sy DeVos and has already made moves to ensure right-wing garbage sites that should be banned from Face­book pure­ly for jour­nal­is­tic integri­ty pur­pos­es will instead be trust­ed con­tent pro­duc­ers and fact-check­ers. And now long-time Fox News vet­er­an Jen­nifer Williams will be work­ing under Brown head­ing up the Face­book News video divi­sion. Because of course.

The impeach­ment of Trump appears to be on course for a quick end fol­low­ing the deci­sion of Sen­ate Repub­li­cans to not call any wit­ness­es and pro­ceed to an acquit­tal vote. The ulti­mate polit­i­cal con­se­quences of acquit­ting Trump with­out call­ing wit­ness­es in the Sen­ate is hard to esti­mate, but it seems like a pret­ty sure bet that the Trump team is going to inter­pret this acquit­tal as a green­light to engage in pret­ty much any polit­i­cal dirty tricks cam­paign it can imag­ine. After all, when Sen­a­tor Lamar Alexan­der — one of the hold out Sen­a­tors who was report­ed­ly on the fence about whether to vote for call­ing wit­ness­es or not — final­ly decid­ed to vote against wit­ness­es late last night, Alexan­der’s rea­son­ing was that House Democ­rats had already proven their case and Trump real­ly did what they accused him of doing but it does­n’t rise to an impeach­able offense so no wit­ness­es were need­ed. So the Repub­li­cans have basi­cal­ly ruled that invit­ing and then extort­ing a for­eign gov­ern­ment to get involved in a US elec­toral dis­in­for­ma­tion cam­paign is accept­able even if they don’t nec­es­sar­i­ly think its fine. It an open invi­ta­tion for not just every Repub­li­can dirty trick imag­in­able but an invi­ta­tion for for­eign gov­ern­ment med­dling too. The Trump pres­i­dent has now become not just the cul­mi­na­tion of Amer­i­ca’s inun­da­tion with dis­in­for­ma­tion but now a val­i­da­tion of it.

So it’s worth not­ing that, days before this deci­sion by the Sen­ate, the Bul­letin of the Atom­ic Sci­en­tists updat­ed the ‘Dooms­day Clock’. It’s now 100 sec­onds from ‘Mid­night’, clos­er than ever. And the explo­sion of dis­in­for­ma­tion cam­paigns and dis­in­for­ma­tion tech­nol­o­gy like ‘deep fakes’ that can send a soci­ety into tur­moil was appar­ent­ly a big part of their rea­son­ing:

ZDNet

The Dooms­day Clock just moved clos­er to mid­night again. Tech is get­ting some of the blame.

Infor­ma­tion war­fare, deep fakes and AI are all adding to the risk of cat­a­stro­phe, sci­en­tists warn.

By Steve Ranger
Jan­u­ary 24, 2020 — 12:37 GMT (04:37 PST)

The Dooms­day Clock has moved clos­er to mid­night than ever before, as sci­en­tists warn that the threats of nuclear war and cat­a­stroph­ic cli­mate change are being com­pound­ed by dig­i­tal infor­ma­tion war­fare, which is mak­ing it hard­er for democ­ra­cies to respond to these dan­gers.

“Human­i­ty con­tin­ues to face two simul­ta­ne­ous exis­ten­tial dangers—nuclear war and cli­mate change—that are com­pound­ed by a threat mul­ti­pli­er, cyber-enabled infor­ma­tion war­fare, that under­cuts soci­ety’s abil­i­ty to respond,” said the Bul­letin of the Atom­ic Sci­en­tists as it moved the Dooms­day Clock from two min­utes to mid­night to 100 sec­onds to mid­night. This shows that they feel the risk of cat­a­stro­phe is greater than ever — even high­er than dur­ing the Cold War.

...

Glob­al insta­bil­i­ty

The group of sci­en­tists warned that sev­er­al major arms con­trol treaties and nego­ti­a­tions have end­ed or been under­mined dur­ing the past year, cre­at­ing an envi­ron­ment con­ducive to a renewed nuclear arms race. They warned that gov­ern­ment action on cli­mate change still falls short.

But they also said that threats to the “infor­ma­tion ecos­phere” — like the spread of mis­in­for­ma­tion and fake news — could also cre­ate dan­ger­ous glob­al insta­bil­i­ty. Ongo­ing dis­in­for­ma­tion cam­paigns are cor­rupt­ing the deci­sion-mak­ing process­es need­ed to tack­le nuclear and cli­mate threats, the sci­en­tists said.

“In the last year, many gov­ern­ments used cyber-enabled dis­in­for­ma­tion cam­paigns to sow dis­trust in insti­tu­tions and among nations, under­min­ing domes­tic and inter­na­tion­al efforts to fos­ter peace and pro­tect the plan­et,” the group said.

While coun­tries have long attempt­ed to use pro­pa­gan­da to dri­ve their par­tic­u­lar polit­i­cal agen­das, the inter­net now pro­vides wide­spread, inex­pen­sive access to world­wide audi­ences. The recent arrival of ‘deep­fake’ audio and video could also under­mine our abil­i­ty to sep­a­rate truth from fic­tion.

“The result­ing false­hoods hold the poten­tial to cre­ate eco­nom­ic, social, and mil­i­tary chaos, increas­ing the pos­si­bil­i­ty of mis­un­der­stand­ings or provo­ca­tions that could lead to war, and foment­ing pub­lic con­fu­sion that leads to inac­tion on seri­ous issues fac­ing the plan­et. Agree­ment on facts is essen­tial to democ­ra­cy and effec­tive col­lec­tive action.”

They sci­en­tists also wor­ry about the impact of arti­fi­cial intel­li­gence and its use in mil­i­tary deci­sion mak­ing and com­mand and con­trol sys­tems.

“The over­all glob­al trend is toward com­plex, high-tech, high­ly auto­mat­ed, high-speed war­fare. The com­put­er­ized and increas­ing­ly AI-assist­ed nature of mil­i­taries, the sophis­ti­ca­tion of their weapons, and the new, more aggres­sive mil­i­tary doc­trines assert­ed by the most heav­i­ly armed coun­tries could result in glob­al cat­a­stro­phe,” the group said.

———–

“The Dooms­day Clock just moved clos­er to mid­night again. Tech is get­ting some of the blame.” by Steve Ranger; ZDNet; 01/24/2020

““Human­i­ty con­tin­ues to face two simul­ta­ne­ous exis­ten­tial dangers—nuclear war and cli­mate change—that are com­pound­ed by a threat mul­ti­pli­er, cyber-enabled infor­ma­tion war­fare, that under­cuts soci­ety’s abil­i­ty to respond,” said the Bul­letin of the Atom­ic Sci­en­tists as it moved the Dooms­day Clock from two min­utes to mid­night to 100 sec­onds to mid­night. This shows that they feel the risk of cat­a­stro­phe is greater than ever — even high­er than dur­ing the Cold War.”

A greater risk of man-made cat­a­stro­phe than dur­ing the Cold War. This is where we are. The rea­sons include ‘oldies’ like the risk of nuclear war. But even there the risks are high­er (thanks in large part to Trump’s shred­ding of nuclear arms treaties). And then there’s the risk of what sound like a ‘Skynet’ sce­nario involv­ing mil­i­taries rely­ing on AI for deci­sion mak­ing and com­mand and con­trol sys­tems:

...
The group of sci­en­tists warned that sev­er­al major arms con­trol treaties and nego­ti­a­tions have end­ed or been under­mined dur­ing the past year, cre­at­ing an envi­ron­ment con­ducive to a renewed nuclear arms race. They warned that gov­ern­ment action on cli­mate change still falls short.

...

They sci­en­tists also wor­ry about the impact of arti­fi­cial intel­li­gence and its use in mil­i­tary deci­sion mak­ing and com­mand and con­trol sys­tems.

“The over­all glob­al trend is toward com­plex, high-tech, high­ly auto­mat­ed, high-speed war­fare. The com­put­er­ized and increas­ing­ly AI-assist­ed nature of mil­i­taries, the sophis­ti­ca­tion of their weapons, and the new, more aggres­sive mil­i­tary doc­trines assert­ed by the most heav­i­ly armed coun­tries could result in glob­al cat­a­stro­phe,” the group said.
...

But it’s the grow­ing threats to the “infor­ma­tion ecos­phere” that runs the risk of dam­ag­ing our abil­i­ty to man­age vir­tu­al­ly every oth­er threat because dis­in­for­ma­tion cam­paigns are already cor­rup­tion the deci­sion-mak­ing process­es need­ed to address all those oth­er threats:

...
But they also said that threats to the “infor­ma­tion ecos­phere” — like the spread of mis­in­for­ma­tion and fake news — could also cre­ate dan­ger­ous glob­al insta­bil­i­ty. Ongo­ing dis­in­for­ma­tion cam­paigns are cor­rupt­ing the deci­sion-mak­ing process­es need­ed to tack­le nuclear and cli­mate threats, the sci­en­tists said.

...

While coun­tries have long attempt­ed to use pro­pa­gan­da to dri­ve their par­tic­u­lar polit­i­cal agen­das, the inter­net now pro­vides wide­spread, inex­pen­sive access to world­wide audi­ences. The recent arrival of ‘deep­fake’ audio and video could also under­mine our abil­i­ty to sep­a­rate truth from fic­tion.
...

And that warn­ing about how dis­in­for­ma­tion threat­ens out col­lec­tive abil­i­ty to deal with ALL OF THE OTHER exis­ten­tial threats is a reminder that sys­tem­at­ic dis­in­for­ma­tion is kind of a meta-exis­ten­tial threat. It lit­er­al­ly makes all oth­er exis­ten­tial threats more like­ly to hap­pen, which arguably makes it the great­est threat of all. If human­i­ty was­n’t so sus­cep­ti­ble to dis­in­for­ma­tion this would­n’t be such a mas­sive threat. But that’s clear­ly not the case. Dis­in­for­ma­tion is win­ning. It real­ly works and is increas­ing­ly cheap and easy to deploy, which is why some­one like Trump can become pres­i­dent and why the far right has been ris­ing across the globe with one big lie cam­paign after anoth­er. And that’s what the Sen­ate Repub­li­cans just rub­ber-stamped and endorsed: the meta-exis­ten­tial threat of sys­tem­at­i­cal­ly trash­ing the infor­ma­tion ecos­phere and the result­ing col­lec­tive insan­i­ty.

Yasha Levine has a short new piece about an inter­est­ing his­tor­i­cal inter­sec­tion between the US’s reha­bil­i­ta­tion of fas­cists and Nazi col­lab­o­ra­tors in the post-WWII era and the coun­terin­sur­gency ori­gins of the devel­op­ment of the inter­net. It’s the kind of his­to­ry that’s long been impor­tant but has sud­den­ly gained a new lev­el of impor­tance now that Pres­i­dent Trump appears to feel ‘unleashed’ fol­low­ing his impeach­ment acquit­tal and will­ing to use the pow­er of his office to pro­tect his friends and attack his polit­i­cal ene­mies:

Levine was giv­en a num­ber of declas­si­fied US Army Counter Intel­li­gence Corp file on Myko­la Lebed. Lebed was one of the many OUN‑B Ukraine fas­cist Nazi col­lab­o­ra­tors who was basi­cal­ly wel­comed into the US’s nation­al secu­ri­ty com­plex and Levine is work­ing on a short biog­ra­phy on him. One par­tic­u­lar file on Lebed was from 1947 and most­ly illeg­i­ble, but it did have a clear stamp at the bot­tom that had the name Col W.P. Yarbor­ough. Yarbor­ough turns out to be a cen­tral fig­ure in the devel­op­ment of the US Army’s spe­cial forces dur­ing this peri­od. Beyond that, he was also a lead­ing fig­ure in the US’s coun­ter­in­tel­li­gence oper­a­tions in the 1960s and it’s in that con­text that Yarbor­ough played a sig­nif­i­cant role in the devel­op­ment of the inter­net’s pre­de­ces­sor, the ARPANET. Levine cov­ered Yarbor­ough’s role in the devel­op­ment of the ARPANET as a coun­terin­sur­gency tool in his book Sur­veil­lance Val­ley. And as he cov­ered in the book, while the coun­terin­sur­gency appli­ca­tions of the orig­i­nal ARPANET was used for the war in Viet­nam, it was also used to com­pile a mas­sive unprece­dent­ed com­put­er­ized data­base on domes­tic polit­i­cal oppo­nents of the war and left-wing groups in gen­er­al.

That’s the main point of Levine’s new piece: the obser­va­tion that the fig­ure who led the devel­op­ment of what was a cut­ting-edge domes­tic sur­veil­lance oper­a­tion pri­mar­i­ly tar­get­ing left-wing polit­i­cal move­ments was also involved with the recruit­ment and uti­liza­tion of WWII fas­cists and Nazis for use in the nation­al secu­ri­ty appa­ra­tus. It’s one of those his­tor­i­cal fun-facts that high­lights how the US’s long-stand­ing ‘anti-com­mu­nism’ agen­da was real­ly an anti-left-wing agen­da that includ­ed the covert sup­pres­sion of domes­tic left-wing move­ments. Fas­cists are fine. Anti-war pro­tes­tors are sub­ver­sives that need to be sur­veilled an ulti­mate­ly neu­tral­ized. It’s a pre­vail­ing theme through­out the Cold War exem­pli­fied by Yarbor­ough’s career. A career that should serve as a warn­ing now that Pres­i­dent Trump appears to feel like he’s been giv­en per­mis­sion to use the full force of the gov­ern­ment to attack per­ceived his polit­i­cal ene­mies:

Yasha.substack.com

From reha­bil­i­tat­ing Nazis to run­ning data-based coun­terin­sur­gency cam­paigns against the Amer­i­can left

The life of a true Amer­i­can hero!

Yasha Levine
02/13/2020

Jared McBride — an amaz­ing his­to­ri­an of nation­al­ism and nation­al­ist iden­ti­ty war­fare in 20th cen­tu­ry Rus­sia, Ukraine, and East­ern Europe — was kind enough to send me the US Army Counter Intel­li­gence Corps’ declas­si­fied file on Myko­la Lebed.

Lebed is not well known out­side a small cir­cle of his­to­ri­ans and jour­nal­ists. But he was an influ­en­tial Ukrain­ian fas­cist and Nazi col­lab­o­ra­tor — a sadist and a mass mur­der­er who after World War II was reha­bil­i­tat­ed and weaponized by the Unit­ed States. He’s an inter­est­ing case and I’ve been try­ing to work up a short biog­ra­phy of him.

I was look­ing through his US Army file when a doc­u­ment from Decem­ber 1947 caught my atten­tion. It had to do with Lebed but was fad­ed and most­ly illeg­i­ble. One thing stood out clear­ly, though. A stamp at the bot­tom that fea­tured a famil­iar name: Col W.P. YARBOROUGH.

From 1945 to 1947, Colonel William P. Yarbor­ough was the Provost Mar­shal of Amer­i­can forces in Vien­na — basi­cal­ly, he was the top mil­i­tary cop of America’s slice of a city that was divid­ed between Britain, France, the Sovi­et Union, and the Unit­ed States. The place, like Berlin, was a spook fest and Yarbor­ough was clear­ly involved in all sorts covert ops.

Why do I know his name?

Turns out Yarbor­ough a tan­gen­tial but very impor­tant fig­ure in the his­to­ry of the ear­ly Inter­net. I wrote a whole sec­tion about him and his work in my book Sur­veil­lance Val­ley.

By the time the 1960s rolled around, Yarbor­ough was regard­ed as an expert on anti-guer­ril­la and coun­terin­sur­gency war­fare. In 1967, while in charge of the U.S. Army’s Intel­li­gence Com­mand, he ini­ti­at­ed a mas­sive, ille­gal domes­tic coun­terin­sur­gency sur­veil­lance pro­gram inside Amer­i­ca that tar­get­ed civ­il rights activists, anti­war pro­test­ers, left­wing stu­dent groups, and any­one who sym­pa­thized with to the oppressed.

It was called CONUS Intel.

Mil­lions of Amer­i­cans got swept up in his para­noid coun­terin­sur­gency pro­gram. Sit­ting sen­a­tors, pro­gres­sive cler­gy, left­ist orga­niz­ers of youth ski clubs, anti­war pro­test­ers, peo­ple who sim­ply attend­ed Mar­tin Luther King’s funer­al — all were spied on and tracked. The scheme explod­ed into a nation­al scan­dal when it was exposed in 1971.

Ulti­mate­ly, the data and sur­veil­lance files that were col­lect­ed by Yarborough’s pro­gram would be dig­i­tized and shared through the ARPANET between the CIA, NSA, FBI, and the White House — as part of an ARPA ini­tia­tive to devel­op dig­i­tal coun­terin­sur­gency tools. We know this hap­pened thanks to the incred­i­ble report­ing of a NBC jour­nal­ist named Ford Rowan.

So Yarbor­ough is con­nect­ed to a big his­tor­i­cal moment: the first known case of the ear­ly Inter­net being used to spy on Amer­i­can civil­ians.

It was fun­ny to see his name on a doc­u­ment con­nect­ed to a Ukrain­ian fas­cist intel­li­gence asset like Myko­la Lebed. From weaponiz­ing Nazi col­lab­o­ra­tors to run­ning com­put­er-based coun­terin­sur­gency cam­paigns against the Amer­i­can left — Yarbor­ough sure lived the life of a true Amer­i­can hero. A lot to be proud of!

It just goes to show how much all of this is linked. The devel­op­ment of coun­terin­sur­gency tech that birthed the Inter­net, the covert reha­bil­i­ta­tion and use of East­ern Euro­pean fas­cists, even the weaponiza­tion of Sovi­et Jews like me — all are part of a larg­er, con­tigu­ous pro­gram of anti-com­mu­nism and Amer­i­can empire expan­sion.

...

————-

“From reha­bil­i­tat­ing Nazis to run­ning data-based coun­terin­sur­gency cam­paigns against the Amer­i­can left” by Yasha Levine; Yasha.substack.com; 02/13/2020

“By the time the 1960s rolled around, Yarbor­ough was regard­ed as an expert on anti-guer­ril­la and coun­terin­sur­gency war­fare. In 1967, while in charge of the U.S. Army’s Intel­li­gence Com­mand, he ini­ti­at­ed a mas­sive, ille­gal domes­tic coun­terin­sur­gency sur­veil­lance pro­gram inside Amer­i­ca that tar­get­ed civ­il rights activists, anti­war pro­test­ers, left­wing stu­dent groups, and any­one who sym­pa­thized with to the oppressed.”

A mas­sive ILLEGAL domes­tic sur­veil­lance oper­a­tion pri­mar­i­ly tar­get­ing the left. That’s what the first ver­sion of the inter­net was used for under the CONUS Intel project. And it was Yarbor­ough — some­one involved with the ear­ly Cold War uti­liza­tion of fas­cists and Nazi — who led that ini­tia­tive:

...
Lebed is not well known out­side a small cir­cle of his­to­ri­ans and jour­nal­ists. But he was an influ­en­tial Ukrain­ian fas­cist and Nazi col­lab­o­ra­tor — a sadist and a mass mur­der­er who after World War II was reha­bil­i­tat­ed and weaponized by the Unit­ed States. He’s an inter­est­ing case and I’ve been try­ing to work up a short biog­ra­phy of him.

I was look­ing through his US Army file when a doc­u­ment from Decem­ber 1947 caught my atten­tion. It had to do with Lebed but was fad­ed and most­ly illeg­i­ble. One thing stood out clear­ly, though. A stamp at the bot­tom that fea­tured a famil­iar name: Col W.P. YARBOROUGH.

From 1945 to 1947, Colonel William P. Yarbor­ough was the Provost Mar­shal of Amer­i­can forces in Vien­na — basi­cal­ly, he was the top mil­i­tary cop of America’s slice of a city that was divid­ed between Britain, France, the Sovi­et Union, and the Unit­ed States. The place, like Berlin, was a spook fest and Yarbor­ough was clear­ly involved in all sorts covert ops.
...

It’s a his­tor­i­cal anec­dote that’s a big reminder that the use state pow­ers to sup­press and min­i­mize left-wing move­ments and indi­vid­u­als is a sig­nif­i­cant part of chap­ter of Amer­i­can his­to­ry that led to where we are today.

It’s worth recall­ing at this point the inter­est­ing sto­ry John Lof­tus had about the white­wash­ing of Myko­la Lebed involv­ing Whitey Bul­ger. It turns out Lebed was cast as an anti-Nazi fight­er in WWII in order to be allowed to get a US visa and become US asset work­ing for the CIA. That white­wash­ing was car­ried about by Dick Sul­li­van, a US Army attor­ney oper­at­ing out of Boston. Lebed was just one of the fas­cists and Nazis who had his back­ground cov­ered up by Sul­li­van. Sul­li­van also hap­pened to be secret mem­ber of Irish Repub­li­can Par­ty (IRA), an alle­giance shared by Bul­ger. Sul­li­van even­tu­al­ly told Bul­ger about an IRA FBI infor­mant, who Bul­ger sub­se­quent­ly killed (this is dis­cussed by Lof­tus on side B of FTR#749).

Now here’s a look at that 1971 NY Times report that ini­tial­ly exposed the Army’s CONUS Intel pro­gram. As the arti­cle describes, while most of the infor­ma­tion fed into this data­base was pro­vid­ed by local police, the FBI, or pub­lic sources, the pro­gram still involved send­ing over 1,000 under­cov­er US Army agents to direct­ly gath­er intel­li­gence. It was only exposed when Sen­a­tor Sam J. Ervin Jr., Demo­c­rat of North Car­oli­na, con­tend­ed that promi­nent polit­i­cal fig­ures in Illi­nois had been under mil­i­tary sur­veil­lance since 1968.

The arti­cle also describes how then-Gen­er­al Yarbor­ough was replaced as the head of CONUS Intel in August of 1968 by Maj. Gen. Joseph McChris­t­ian. AFter McChris­t­ian was briefed on the pro­gram he imme­di­ate­ly asked his sub­or­di­nates for ways to cut it back. But McChris­t­ian ran into resis­tance from the “domes­tic war room” and oth­er gov­ern­ment agen­cies, par­tic­u­lar­ly the Jus­tice Depart­ment, which said it need­ed this domes­tic intel­li­gence. All in all, the CONUS Intel chap­ter of Amer­i­can his­to­ry is a chap­ter that’s become omi­nous­ly rel­e­vant for the age of ‘Trump unleashed’:

The New York Times

Army Spied on 18,000 Civil­ians in 2‑Year Oper­a­tion

By Richard Hal­lo­ran Spe­cial to The New York Times
Jan. 18, 1971

WASHINGTON, Jan. 17—The Unit­ed States Army fed the names of about 18,000 Amer­i­can civil­ians into its com­put­ers, dossiers and files in a wide rang­ing intel­li­gence oper­a­tion dur­ing the tumul­tuous days of civ­il dis­tur­bances from the sum­mer of 1967 through the fall of 1969.

In the oper­a­tion, which was ordered end­ed last year, 1,000 Army agents gath­ered per­son­al and polit­i­cal infor­ma­tion on obscure per­sons, as well as the promi­nent, on advo­cates of vio­lent protest arid par­tic­i­pants in legit­i­mate polit­i­cal activ­i­ty, on the Nation­al Asso­ci­a­tion for the Advance­ment of Col­ored Peo­ple and the John Birch Soci­ety, on the Black Pan­thers and the Ku Klux Klan, on the Stu­dents for a Demo­c­ra­t­ic Soci­ety and the Daugh­ters of the Amer­i­can Rev­o­lu­tion. The empha­sis was on rad­i­cals, black mil­i­tants and dis­senters against the war in Viet­nam.

The mil­i­tary intel­li­gence oper­a­tion picked up much of its infor­ma­tion from local police offi­cials and the Fed­er­al Bureau of Inves­ti­ga­tion, but sup­ple­ment­ed that data and col­lect­ed its own through agents pos­ing as mem­bers of the groups under sur­veil­lance, as news­men, or mere­ly as inter est­ed bystanders.

Thus, a black agent reg­is­tered at New York Uni­ver­si­ty in 1968 to report on stu­dents tak­ing a course in black stud­ies. An oth­er agent joined the Youth Inter­na­tion­al par­ty, or Yip­pies, and slept along­side its can­di­date, a pig named “Piga­sus,” dur­ing the counter-inau­gur­al demon­stra­tion here in Jan­u­ary, 1969.

The Army now autho­rizes only lim­it­ed intel­li­gence gath­er­ing on inci­dents that might lead to a Pres­i­den­tial call for Fed­er­al troops. But atten­tion was sharply focused last month on the Army oper­a­tion when Sen­a­tor Sam J. Ervin Jr., Demo­c­rat of North Car­oli­na, con­tend­ed that promi­nent polit­i­cal fig­ures in Illi­nois had been under mil­i­tary sur­veil­lance since 1968.

Sen­a­tor Ervin is skep­ti­cal of the Army’s announce­ment about halt­ing the spy­ing and has sched­uled hear­ings by the Sub­com­mit­tee on Con­sti­tu­tion­al Rights, of which he is chair­man, to begin Feb. 23. He said the Army “must dis­close in full what hap­pened and why it hap­pened and what has been done to insure that it will nev­er hap­pen again.”

Details of the oper­a­tion, known as Con­ti­nen­tal Unit­ed States Intel­li­gence, or Conus Intel, emerged from inter­views with civil­ian and uni­formed Pen­ta­gon offi­cials, Con­gres­sion­al sources, agents of the Secret Ser­vice and for­mer agents, and from a study of Army doc­u­ments and files.

The find­ings includ­ed the fol­low­ing:

¶Direc­tives from Cab­i­net-lev­el offi­cials, autho­riz­ing intel­li­gence gath­er­ing to help the Army car­ry out its mis­sion of quelling civ­il dis­or­ders, were impre­cise. Army guide­lines for sub­or­di­nate com­mands were loose­ly drawn—like “a license to steal,” one Pen­ta­gon source said.

¶In a vari­a­tion of an old Army game, each sub­or­di­nate expand­ed on his instruc­tions to please his supe­ri­ors and to pro­tect him­self from charges that he had not done his job.

¶Once start­ed, the intel­li­gence oper­a­tion gen­er­at­ed a demand for its prod­uct from the Jus­tice Depart­ment, the F.B.I., police depart­ments and oth­er gov­ern­ment agen­cies. A source close to the oper­a­tion said, “We cre­at­ed addicts for this stuff all over the Gov­ern­ment.”

¶Some younger agents enjoyed play­ing James Bond. Large­ly col­lege-edu­cat­ed and work­ing away from reg­u­lar Army dis­ci­pline, these men found it more fun to spy on polit­i­cal agi­ta­tors than to make the rou­tine secu­ri­ty checks that have long been a pri­ma­ry task of mil­i­tary intel­li­gence.

¶Conus Intel was but one part of a vast, inter­lock­ing intel­li­gence exchange that Pres­i­dents Kennedy and John­son, and prob­a­bly Pres­i­dent Nixon, knew was in oper­a­tion, although they may not have been aware of all of its details.

¶There was no con­spir­a­cy, as far as could be dis­cerned, by the mil­i­tary to sub­vert polit­i­cal lib­er­ties. One crit­i­cal for­mer agent said that “these were not malev­o­lent men.” Rather, he said, they were well-inten­tioned men car­ry­ing out what they con­sid­ered to be legit­i­mate orders from polit­i­cal author­i­ties.

Com­ment­ing on this last point, the Army’s gen­er­al coun­sel, Robert E. Jor­dan 3d, said, “I hon­est­ly believe we drift­ed into this area with­out quite real­iz­ing what we were get­ting into and because no one else was around to do the job.” He added:

“I’m con­vinced that no one intend­ed to spy on indi­vid­u­als or con­trol civil­ian life in any way. But I also believe that some of the things begun, if expand­ed, sure as hell posed a real risk.”

Over­hauled in 1963

The domes­tic mil­i­tary intel­li­gence appa­ra­tus, which was first involved in a civ­il dis­tur­bance dur­ing the riots at Oxford, Miss., in 1962, was over­hauled when a delayed secu­ri­ty check in 1963 showed that an Army sergeant in a sen­si­tive post had been a Sovi­et agent. That led to the for­ma­tion, on Jan. 1, 1965, of the Army Intel ligence Com­mand at Fort Holabird Md.

The eight mil­i­tary intel­li­gence groups around the coun­try, each with about 400 men, were trans­ferred from area com­man­ders to the cen­tral­ized con­trol of the Army Intel­li­gence Com­mand to make secu­ri­ty clear­ances and oth­er anti sub­ver­sive oper­a­tions more effi­cient.

That set up the appa­ra­tus for the sub­se­quent col­lec­tion of in for­ma­tion from the 1,000 agents in the 300 mil­i­tary intel­li­gence field offices across the nation. The intel­li­gence was ana­lyzed by the Coun­ter­in­tel­li­gence Analy­sis Detach­ment, or CIAD, in the office of the Army’s assis­tant chief of staff for intel­li­gence.

Riots and Protests

Dur­ing the sum­mer and fall of 1965, the nation was shak­en by racial riots in the Watts sec­tion of Los Ange­les and else­where, and by the first protests against the increas­ing Amer­i­can involve­ment in Viet­nam. Fed­er­al troops were not called to curb the riots and protests, but it became evi­dent that they might be need­ed.

In 1966, the Army Intel­li­gence Com­mand instruct­ed the mil­i­tary intel­li­gence groups to col­lect basic infor­ma­tion about cities that might be use­ful if the Army were called.

Not much was done about gath­er­ing the infor­ma­tion, but agents mak­ing rou­tine vis­its to cam­pus­es for back­ground in ves­ti­ga­tions began pick­ing up leaflets from anti­war dis­senters and lis­ten­ing in on their ral­lies. The Coun­ter­in­tel­li­gence Analy­sis Detach­ment start­ed mon­i­tor­ing expres­sions of dis­sent and black mil­i­tance, most­ly by hav­ing a few men clip news­pa­pers. Agents in an unmarked truck fol­lowed James Mered­ith on his “walk against fear” through Mis­sis­sip­pi.

Caught Unpre­pared

In 1967, the Army was caught unpre­pared when racial riots broke out in Newark and Detroit. Army troops called in to help restore order had lit­tle more than Esso road maps to guide them in both cities.

The Army’s chief intel­li­gence offi­cer then was Maj. Gen. William P. Yarbor­ough, a long time coun­ter­in­tel­li­gence and psy­cho­log­i­cal war­fare spe­cial ist. The flam­boy­ant gen­er­al, known as “Big Y” for the way he signed mem­o­ran­dums, told sub­or­di­nates that the riot­ers were “insur­gents” manip­u­lat­ed by the Communists—and he began try­ing to find out more about them.

Gen­er­al Yarbor­ough, now a lieu­tenant gen­er­al serv­ing in. Hawaii, said last week through a Pen­ta­gon spokesman that “my rec­om­men­da­tion that Unit­ed States Army plan­ners use the coun­terin­sur­gency plan­ning guide in con­nec­tion with mas­sive civ­il dis­tur­bances inside the Unit­ed States did not in any way imply that I believed those phe­nom­e­na con­sti­tut­ed actu­al insur­gency.”

Gen­er­al Yarbor­ough ordered a Conus Intel com­mu­ni­ca­tions cen­ter known as “Oper­a­tions IV” to be set up at Fort Holabird and a nation­wide tele­type net­work that would feed infor­ma­tion to it. Large amounts of infor­ma­tion came from the F.B.I. and local police depart ments, but he also instruct­ed mil­i­tary intel­li­gence agents to pick up infor­ma­tion on their own.

Con­ti­nen­tal Unit­ed States Intel­li­gence paid par­tic­u­lar atten­tion to the well-pub­li­cized plans for the anti-Viet­nam march on the Pen­ta­gon in Octo­ber, 1967. Agents from the New York field office of the 108th Mil­i­tary Intel­li­gence Group, for exam­ple, rode bus­es into Wash­ing­ton and stayed with the crowd all through the demon­stra­tion.

But the Army under­es­ti­mat­ed the num­ber of peo­ple that would show up, how long they would stay, and the degree of vio­lence they would attempt. For those fail­ures, senior offi­cers caught what one source described as “undi­lut­ed hell” from high polit­i­cal lead­ers, appar­ent­ly includ­ing Pres­i­dent John­son.

Review Urged

Imme­di­ate­ly after the march on the Pen­ta­gon, Sec­re­tary of Defense Robert S. McNa­ma­ra asked the Under Sec­re­tary of the Army, David E. McGif­fert, to review the entire role of Fed­er­al troops in civ­il dis­tur­bances. Mr. McGif­fert called meet­ing that includ­ed War­ren Christo­pher, the Deputy Attor­ney Gen­er­al; Stephen Pol­lak, spe­cial assis­tant to the Pres­i­dent, and numer­ous oth­ers from the Depart­ments of Defense and Jus­tice, the F.B.I., the Secret Ser­vice and local police offi­cials.

Out of their study came the Army’s civ­il dis­tur­bance plan in Decem­ber, 1967. Two months lat­er, an intel­li­gence annex that set out infor­ma­tion require­ments for Army field com­man­ders was added to the plan. That was the begin­ning of the “city books” that detailed the infor­ma­tion a com­man­der might need if he moved troops into an urban area.

Much of the infor­ma­tion involved tac­ti­cal intel­li­gence— where troops would land, where they would bivouac, where the hos­pi­tals and the police sta­tions were sit­u­at­ed. Army offi­cers met with the police offi­cials to see where trou­ble might occur. They talked with police offi­cers down to the precinct lev­el to spot gun shops and liquor stores that might be tar­gets for riot­ers.

Pos­si­ble Agi­ta­tors

In addi­tion, Army offi­cers slid into the polit­i­cal sphere by ask­ing the police for the names and pic­tures of pos­si­ble riot agi­ta­tors. They also asked the police for the names and pic­tures of peo­ple who might be will­ing to help calm a crowd.

That infor­ma­tion, along with oth­er mate­r­i­al from the F.B.I. and the Secret Ser­vice, was fed back to Wash­ing­ton, where it went into the com­pendi­um” com­piled by the Coun­ter­in­tel­li­gence Analy­sis Detach­ment. “The com­pendi­um” was a two-vol­ume ency­clo­pe­dia that con ained pic­tures and data, includ­ing the polit­i­cal beliefs, on peo­ple who might either foment or help stop a civ­il dis­tur­bance. The coun­ter­in­tel­li­gence group was also charged by Mr. Mc Gif­fert with try­ing to pre­dict when and where a civ­il dis­tur­bance might break out.

But the assas­si­na­tion of the Rev. Dr. Mar­tin Luther King Jr., in April, 1968, put an end to that idea. The riot­ing that occurred in 100 cities after his assas­si­na­tion showed that the site of a civ­il dis­tur­bance could not be pre­dict­ed.

High-Lev­el Review

Although the Army was bet­ter pre­pared to han­dle the dis­or­ders in Wash­ing­ton, Balti more and Chica­go than it had been dur­ing ear­li­er riots in Newark and Detroit, the need for Fed­er­al troops and the nation­wide ten­sion stim­u­lat­ed an oth­er high-lev­el review. At meet­ings in the Pen­ta­gon on April 12 and in the White House on April 15, 1968, Mr. McGif­fert pro­posed that Army intel­li­gence con­cen­trate on civ­il dis­tur­bance warn­ings.

Out of those meet­ings also came a require­ment that the Army be pre­pared to send 10,000 troops on short notice to any one of 25 cities. That num­ber was lat­er reduced to about 10 cities where the Nation­al Guard and the local police were con­sid­ered unable to han­dle things on their own.

Through the sum­mer of 1968, Army intel­li­gence oper­a­tions inten­si­fied. The Army put into effect its civ­il dis­tur­bance in for­ma­tion plan on May 2, giv­ing its agents more col­lec­tion require­ments.

They were told to report on every­thing that bore there motest con­nec­tion to civ­il dis­tur­bances. Maj. Gen. William H. Blake­field, the intel­li­gence com­man­der, told his sub­or­di­nates to “beat the Asso­ci­at­ed Press” in their report­ing.

Gen­er­al Yarbor­ough set up a task force in the Coun­ter­in­tel­li­gence Analy­sis Detach­ment to study infor­ma­tion about the 1968 poor peo­ple’s cam­paign and Res­ur­rec­tion City in Wash­ing­ton, which were close­ly scru­ti­nized by mil­i­tary intel­li­gence agents.

The intel­li­gence com­mand start­ed dis­trib­ut­ing its “black list,” which includ­ed names, pic­tures, per­son­al data and polit­i­cal char­ac­ter­i­za­tions, such as “rad­i­cal” or “mil­i­tant,” of poten­tial trou­ble­mak­ers. The “black­list” went to law enforce­ment agen­cies at all lev­els, as well as Army com­man­ders and mil­i­tary intel­li­gence groups.

In June, 1968, Sen­a­tor Robert F. Kennedy was assas­si­nat­ed and Con­gress passed a res­o­lu­tion giv­ing the Secret Ser­vice the author­i­ty to draw on the Army and oth­er Fed­er­al agen­cies for help in pro­tect­ing nation­al polit­i­cal can­di­dates.

Paul Nitze, the Deputy Sec­re­tary of Defense, signed an order on June 8 that was the most explic­it direc­tive until then on the Army’s intel­li­gence gath­er­ing pro­ce­dures. The order gave for­mal instruc­tions to pro­vide to the Pen­ta­gon all of the essen­tial intel­li­gence data on civ­il dis­tur­bances.

The intel­li­gence com­mand at Fort Holabird began using com­put­ers to store infor­ma­tion on civ­il dis­tur­bances. One data bank con­tained a file on inci­dents, a sec­ond a bio­graph­i­cal file on sol­diers who were con sidered pos­si­ble dis­senters.

A sim­i­lar data bank was opened at the Con­ti­nen­tal Army Com­mand head­quar­ters at Fort Mon­roe, Va., for a pro­gram called Rita, for Resis­tance in the Army. Still anoth­er data bank was at III Corps head quar­ters at Fort Hood, Tex. This data bank con­cen­trat­ed on civ­il dis­tur­bance infor­ma­tion because two Army divi­sions at Fort Hood had antiri­ot respon sibil­i­ties.

‘Domes­tic War Room’

The Direc­torate for Civ­il Dis­tur­bance Plan­ning and Oper­a­tions was set up in June, 1968, in what came to be known as the “domes­tic war room” in the base­ment of the Pen­ta­gon. This group was respon­si­ble for order­ing air­lifts, troop deploy­ment and logis­tics in a civ­il dis­or­der and became a major con­sumer of Conus Intel’s in for­ma­tion.

When the Repub­li­cans con­vened in Mia­mi in July, 1968, to nom­i­nate Mr. Nixon as their Pres­i­den­tial can­di­date, the Air Force was in charge of the Defense Depart­men­t’s role there. The Army, how­ev­er, fur­nished about 30 men from the Crim­i­nal Inves­ti­ga­tion Divi­sion of the mil­i­tary police, plus 17 dog han­dlers and 40 bomb dis­pos­al spe­cial­ists, to pro­tect the can­di­dates and the del­e­gates.

Mil­i­tary intel­li­gence agents from the 111th Group at Fort McPher­son, Ga., were in Mia­mi to watch for civ­il dis­tur­bances. Most of the agents were post­ed out­side the con­ven­tion hall and in Lib­er­ty City, near Mia­mi, where racial dis­or­ders occurred.

Min­gled With Del­e­gates

But there was also an intel­li­gence com­mand post inside the hall. Agents were sta­tioned around the edge of the floor, and sev­er­al offi­cers in civil­ian clothes min­gled with the del­e­gates. No polit­i­cal infor­ma­tion, how­ev­er, appeared to have been col­lect­ed.

At the Demo­c­ra­t­ic Nation­al Con­ven­tion in Chica­go the next month, the Army again sent mil­i­tary police in civil­ian clothes to help the Secret Ser­vice pro­tect the can­di­dates. Intel­li­gence agents from the 113th Group, con­sid­ered among the most effec­tive, report­ed on civ­il dis­tur­bances to inform the 7,000 troops posi­tioned near the city.

In addi­tion, elec­tron­ic spe­cial­ists from the Army Secu­ri­ty Agency inter­cept­ed radio mes­sages trans­mit­ted on walkie-talkies used by lead­ers of the anti-Viet­nam demon­stra­tors. Pen­ta­gon offi­cials adamant­ly assert­ed that no tele­phones were tapped or rooms bugged.

Cov­er Orga­ni­za­tion

An intel­li­gence crew of cam­era­men, pos­ing as news­men from a cov­er orga­ni­za­tion called Mid-West Video News, took pic­tures of the demon­stra­tors and obtained a filmed inter­view with Abbie Hoff­man, who lat­er was one of the defen­dants at the tri­al of the Chica­go Sev­en.

By the end of 1968, the Army intel­li­gence oper­a­tion was mov­ing at top speed. When dis­senters planned their counter-inau­gur­al demon­stra­tions in Wash­ing­ton in Jan­u­ary, 1969, the Army knew how many pro­test­ers would show up and what they planned to do.

Through­out 1969, Army intel­li­gence turned out an aver­age of 1,200 spot reports each month on inci­dents around the nation. By that time, there were exten­sive inci­dent and per­son­al­i­ty files in every mil­i­tary intel­li­gence field, region­al and group head­quar­ters, plus the com­put­er banks at Fort Holabird, Fort Mon­roe and Fort Hood. In addi­tion, the counter-intel­li­gence detach­men­t’s 120,000 pages of micro­film con tained about 5,000 pages on civil­ians.

At one mil­i­tary intel­li­gence group, a file was opened on the D.A.R. When a man rep­re­sent­ing him­self as an offi­cial of the orga­ni­za­tion asked the Army for a senior offi­cer as a speak­er, mil­i­tary intel­li­gence was asked to check the D.A.R. to see whether it had male employes. It did.

Just how exten­sive all those files were, nobody knows pre­cise­ly. The main com­put­er was pro­gramed for inci­dents rather than peo­ple. Not all of the files were ever com­piled in one place to elim­i­nate dupli­ca­tion. More over, the Army says most of them have been destroyed by now and those that remain have been sealed for pos­si­ble use in appeals to suits brought by the Amer­i­can Civ­il Lib­er­ties Union.

Laws of Physics

Even as the Army intel­li­gence oper­a­tion was speed­ing along, how­ev­er, some efforts were begin­ning to be made to slow it down. But stop­ping it proved dif­fi­cult. Bureau­cra­cies seem to fol­low the laws of physics—a bureau­cra­cy at rest tends to stay at rest; a bureau­cra­cy in motion tends to remain in motion.

Gen­er­al Yarbor­ough, who had start­ed the Army’s domes­tic intel­li­gence oper­a­tions, was replaced in August, 1968, by Maj. Gen. Joseph A. McChris­t­ian, a for­mer head of all mil­i­tary intel­li­gence in Viet­nam.

In tak­ing over his new as sign­ment, Gen­er­al McChris­t­ian was briefed on Con­ti­nen­tal Unit­ed States Intel­li­gence, and imme­di­ate­ly asked his sub­or­di­nates to find ways to cut it back. He was pri­mar­i­ly con cerned with the time it was tak­ing away from oth­er tasks in mil­i­tary intel­li­gence.

But the gen­er­al ran into resis­tance from the “domes­tic war room” and oth­er Gov­ern ment agen­cies, par­tic­u­lar­ly the Jus­tice Depart­ment, that said they need­ed the infor­ma­tion com­ing from the intel­li­gence oper­a­tion.

Request for Film

The Under Sec­re­tary of the Army, Mr. McGif­fert, start­ed to won­der, how­ev­er, about the pro­pri­ety of the oper­a­tion in Octo­ber, when he dis­cov­ered that agents had filmed a demon­stra­tion dur­ing the Demo­c­ra­t­ic Nation­al Con­ven­tion in Chica­go. That came to his atten­tion when the Jus­tice De part­ment asked him for the video tape for pos­si­ble use in the Chica­go Sev­en tri­al.

In a mem­o­ran­dum dat­ed Feb. 5, 1969, short­ly before he left the Gov­ern­ment after the Nixon Admin­is­tra­tion took over, Mr. McGif­fert said that mil­i­tary intel­li­gence might be in dan­ger of exceed­ing its author­i­ty and that hence­forth no covert oper­a­tions would be con­duct­ed.

Short­ly after, the intel­li­gence com­mand stopped dis­trib­ut­ing its “black­list,” but kept it up to date until the end of 1969, when it was ordered with­drawn. Because so many copies had been sent out, the Pen­ta­gon could not be sure that they were all returned for destruc­tion.

Agree­ment Sought

About the same time, Mr. Jor­dan, the Army gen­er­al coun­sel, began explor­ing with Deputy Attor­ney Gen­er­al Richard G. Klein­di­enst the pos­si­bil­i­ty of hav­ing the Jus­tice Depart­ment take over intel­li­gence gath­er­ing on civ­il dis­tur­bances. But at a meet­ing on April 1, he was unable to obtain an agree­ment. Mr. Klein­di­enst con­tend­ed that his depart­ment lacked the man­pow­er to do the job.

Nev­er­the­less, Gen­er­al McChris­t­ian ordered some restric­tions on his own author­i­ty, instruct­ing that more time be put on secu­ri­ty clear­ances and oth­er tasks of pro­tect­ing Army instal­la­tions.

That is appar­ent­ly where the mat­ter stood until last Jan­u­ary, when a for­mer cap­tain of mil­i­tary intel­li­gence, Christo­pher H. Pyle, pub­lished a long arti­cle in The Wash­ing­ton Month­ly describ­ing some of the oper­a­tions of Con­ti­nen­tal Unit­ed States Intel­li­gence. He crit­i­cized the Army for going beyond the needs for infor­ma­tion on civ­il dis­tur­bances.

In response, Gen­er­al McChris­t­ian instruct­ed Gen­er­al Blake­field to exam­ine all pro­ce­dures in the intel­li­gence com­mand that might threat­en polit­i­cal free­dom and ordered the head of coun­ter­in­tel­li­gence in the Pen­ta­gon, Col. John W. Down­ie, to do the same with pol­i­cy direc­tives.

Print­out on Mrs. King

Mr. Jor­dan, the gen­er­al coun­sel, then went to Fort Holabird to exam­ine the com­put­er data bank. He asked for a print­out on sev­er­al names, includ­ing that of Mrs. Coret­ta King, Dr. King’s wid­ow. The print­out showed the pos­si­bil­i­ties for using the data bank to check on peo­ple rather than mere­ly inci­dents.

Last Feb. 19, there­fore, Gen­er­al McChris­t­ian ordered that both data banks at Fort Hola bird be destroyed. The data banks at Fort Hood and Fort Mon­roe were destroyed lat­er. One print­out from each was kept for pos­si­ble court pro­ceed­ings.

Mean­while, Mr. Pyle’s arti­cle prompt­ed sev­er­al Con­gres­sion­al inquiries, includ­ing one from Sen­a­tor Ervin. Mr. Jor­dan, in a let­ter to the Sen­a­tor on Feb. 25, said the Army had restrict­ed report­ing “to inci­dents which may be beyond the capa­bil­i­ty of local and state author­i­ties to con­trol and may require the deploy­ment of Fed­er­al troops.”

Thad­deus R. Beal, the Under Sec­re­tary of the Army, said in anoth­er let­ter to Sen­a­tor Ervin on March 20 that spot reports on out­breaks of vio­lence would con­tin­ue but that they would be kept only for 60 days.

Three Army direc­tives, last April 1, June 9 and Dec, 15, increas­ing­ly tight­ened the restric­tions on col­lect­ing and report­ing infor­ma­tion and on the use of data banks. They flat­ly pro­hib­it­ed the use of com­put­ers to store infor­ma­tion on civil­ians.

Last sum­mer, Colonel Down­ie, the Pen­tagon’s coun­ter­in­tel­li­gence chief, vis­it­ed var­i­ous group head­quar­ters and field offices to see that files on civil­ians were destroyed. For­mer agents said, how­ev­er, that some agents had evad­ed the order by hid­ing files or tak­ing them home as “per­son­al” papers. Gen­er­al McChris­t­ian also ordered the coun­ter­in­tel­li­gence detach ment to cleanse its micro­film of infor­ma­tion on civil­ians.

Things were qui­et then until last month, when Sen­a­tor Ervin opened the issue up again by charg­ing that Con­ti­nen­tal Unit­ed States intel­li­gence had spied on Sen­a­tor Adlai E. Steven­son 3d, Rep­re­sen­ta­tive Abn­er Mik­va, and for­mer Gov. Otto Kern­er, all of Illi­nois.

Sen­a­tor Erv­in’s alle­ga­tions were based on state­ments by John M. O’Brien, a for­mer staff sergeant who had served with the 113th Group in Chica­go. He made them again under oath in a suit brought by the A.C.L.U. in Chica­go. But Sec­re­tary Resor cat­e­gor­i­cal­ly denied the charges.

....

———–

“Army Spied on 18,000 Civil­ians in 2‑Year Oper­a­tion” by Richard Hal­lo­ran; The New York Times; 01/18/1971

“In the oper­a­tion, which was ordered end­ed last year, 1,000 Army agents gath­ered per­son­al and polit­i­cal infor­ma­tion on obscure per­sons, as well as the promi­nent, on advo­cates of vio­lent protest arid par­tic­i­pants in legit­i­mate polit­i­cal activ­i­ty, on the Nation­al Asso­ci­a­tion for the Advance­ment of Col­ored Peo­ple and the John Birch Soci­ety, on the Black Pan­thers and the Ku Klux Klan, on the Stu­dents for a Demo­c­ra­t­ic Soci­ety and the Daugh­ters of the Amer­i­can Rev­o­lu­tion. The empha­sis was on rad­i­cals, black mil­i­tants and dis­senters against the war in Viet­nam.”

1,000 Army agents col­lect­ing domes­tic intel­li­gence. Some­times pos­ing as mem­bers of the groups under sur­veil­lance, or mem­bers of the press, or just ran­dom bystanders. It’s kind of a night­mare sit­u­a­tion from a con­sti­tu­tion­al per­spec­tive:

...
The mil­i­tary intel­li­gence oper­a­tion picked up much of its infor­ma­tion from local police offi­cials and the Fed­er­al Bureau of Inves­ti­ga­tion, but sup­ple­ment­ed that data and col­lect­ed its own through agents pos­ing as mem­bers of the groups under sur­veil­lance, as news­men, or mere­ly as inter­est­ed bystanders.

Thus, a black agent reg­is­tered at New York Uni­ver­si­ty in 1968 to report on stu­dents tak­ing a course in black stud­ies. An oth­er agent joined the Youth Inter­na­tion­al par­ty, or Yip­pies, and slept along­side its can­di­date, a pig named “Piga­sus,” dur­ing the counter-inau­gur­al demon­stra­tion here in Jan­u­ary, 1969.
...

The pro­gram emerged from cre­ation of the Army Intel­li­gence Com­mand at Fort Holabird Md. in 1965 that con­nect­ed 300 mil­i­tary intel­li­gence field offices across the US:

...
Over­hauled in 1963

The domes­tic mil­i­tary intel­li­gence appa­ra­tus, which was first involved in a civ­il dis­tur­bance dur­ing the riots at Oxford, Miss., in 1962, was over­hauled when a delayed secu­ri­ty check in 1963 showed that an Army sergeant in a sen­si­tive post had been a Sovi­et agent. That led to the for­ma­tion, on Jan. 1, 1965, of the Army Intel­li­gence Com­mand at Fort Holabird Md.

The eight mil­i­tary intel­li­gence groups around the coun­try, each with about 400 men, were trans­ferred from area com­man­ders to the cen­tral­ized con­trol of the Army Intel­li­gence Com­mand to make secu­ri­ty clear­ances and oth­er anti sub­ver­sive oper­a­tions more effi­cient.

That set up the appa­ra­tus for the sub­se­quent col­lec­tion of in for­ma­tion from the 1,000 agents in the 300 mil­i­tary intel­li­gence field offices across the nation. The intel­li­gence was ana­lyzed by the Coun­ter­in­tel­li­gence Analy­sis Detach­ment, or CIAD, in the office of the Army’s assis­tant chief of staff for intel­li­gence.
...

Then in 1966, fol­low­ing the race riots of 1965 and the first protests against the US war in Viet­nam, when fed­er­al troops were called in, the Army Intel­li­gence Com­mand instruct­ed those mil­i­tary intel­li­gence offices to start col­lect­ing infor­ma­tion that might be use­ful if the Army was called into a city. A side effect of this order was agents mak­ing reg­u­lar vis­its to cam­pus­es and col­lect­ing anti-war lit­er­a­ture. This result­ed in the Coun­ter­in­tel­li­gence Analy­sis Detach­ment mon­i­tor­ing expres­sions of dis­sent and black mil­i­tants:

...
Riots and Protests

Dur­ing the sum­mer and fall of 1965, the nation was shak­en by racial riots in the Watts sec­tion of Los Ange­les and else­where, and by the first protests against the increas­ing Amer­i­can involve­ment in Viet­nam. Fed­er­al troops were not called to curb the riots and protests, but it became evi­dent that they might be need­ed.

In 1966, the Army Intel­li­gence Com­mand instruct­ed the mil­i­tary intel­li­gence groups to col­lect basic infor­ma­tion about cities that might be use­ful if the Army were called.

Not much was done about gath­er­ing the infor­ma­tion, but agents mak­ing rou­tine vis­its to cam­pus­es for back­ground in ves­ti­ga­tions began pick­ing up leaflets from anti­war dis­senters and lis­ten­ing in on their ral­lies. The Coun­ter­in­tel­li­gence Analy­sis Detach­ment start­ed mon­i­tor­ing expres­sions of dis­sent and black mil­i­tance, most­ly by hav­ing a few men clip news­pa­pers. Agents in an unmarked truck fol­lowed James Mered­ith on his “walk against fear” through Mis­sis­sip­pi.
...

After race riots broke out in Newark and Detroit in 1967, Gen­er­al Yarbor­ough ordered a Conus Intel com­mu­ni­ca­tions cen­ter known as “Oper­a­tions IV” to be set up at Fort Holabird and a nation­wide tele­type net­work that would feed infor­ma­tion to it. It was this ear­ly telecom­mu­ni­ca­tion infra­struc­ture that allowed for col­lec­tion of infor­ma­tion from around the coun­try that we now know was the ear­ly incar­na­tion of the inter­net:

...
Caught Unpre­pared

In 1967, the Army was caught unpre­pared when racial riots broke out in Newark and Detroit. Army troops called in to help restore order had lit­tle more than Esso road maps to guide them in both cities.

The Army’s chief intel­li­gence offi­cer then was Maj. Gen. William P. Yarbor­ough, a long time coun­ter­in­tel­li­gence and psy­cho­log­i­cal war­fare spe­cial ist. The flam­boy­ant gen­er­al, known as “Big Y” for the way he signed mem­o­ran­dums, told sub­or­di­nates that the riot­ers were “insur­gents” manip­u­lat­ed by the Communists—and he began try­ing to find out more about them.

Gen­er­al Yarbor­ough, now a lieu­tenant gen­er­al serv­ing in. Hawaii, said last week through a Pen­ta­gon spokesman that “my rec­om­men­da­tion that Unit­ed States Army plan­ners use the coun­terin­sur­gency plan­ning guide in con­nec­tion with mas­sive civ­il dis­tur­bances inside the Unit­ed States did not in any way imply that I believed those phe­nom­e­na con­sti­tut­ed actu­al insur­gency.”

Gen­er­al Yarbor­ough ordered a Conus Intel com­mu­ni­ca­tions cen­ter known as “Oper­a­tions IV” to be set up at Fort Holabird and a nation­wide tele­type net­work that would feed infor­ma­tion to it. Large amounts of infor­ma­tion came from the F.B.I. and local police depart ments, but he also instruct­ed mil­i­tary intel­li­gence agents to pick up infor­ma­tion on their own.

Con­ti­nen­tal Unit­ed States Intel­li­gence paid par­tic­u­lar atten­tion to the well-pub­li­cized plans for the anti-Viet­nam march on the Pen­ta­gon in Octo­ber, 1967. Agents from the New York field office of the 108th Mil­i­tary Intel­li­gence Group, for exam­ple, rode bus­es into Wash­ing­ton and stayed with the crowd all through the demon­stra­tion.

But the Army under­es­ti­mat­ed the num­ber of peo­ple that would show up, how long they would stay, and the degree of vio­lence they would attempt. For those fail­ures, senior offi­cers caught what one source described as “undi­lut­ed hell” from high polit­i­cal lead­ers, appar­ent­ly includ­ing Pres­i­dent John­son.
...

Fol­low­ing a mas­sive anti-war march on the Pen­ta­gon in 1967, a review of the role of fed­er­al troops in civ­il dis­tur­bances was set up, lead­ing to “city books” plans that detailed how a mil­i­tary com­man­der might need to move troops into an urban area:

...
Review Urged

Imme­di­ate­ly after the march on the Pen­ta­gon, Sec­re­tary of Defense Robert S. McNa­ma­ra asked the Under Sec­re­tary of the Army, David E. McGif­fert, to review the entire role of Fed­er­al troops in civ­il dis­tur­bances. Mr. McGif­fert called meet­ing that includ­ed War­ren Christo­pher, the Deputy Attor­ney Gen­er­al; Stephen Pol­lak, spe­cial assis­tant to the Pres­i­dent, and numer­ous oth­ers from the Depart­ments of Defense and Jus­tice, the F.B.I., the Secret Ser­vice and local police offi­cials.

Out of their study came the Army’s civ­il dis­tur­bance plan in Decem­ber, 1967. Two months lat­er, an intel­li­gence annex that set out infor­ma­tion require­ments for Army field com­man­ders was added to the plan. That was the begin­ning of the “city books” that detailed the infor­ma­tion a com­man­der might need if he moved troops into an urban area.

Much of the infor­ma­tion involved tac­ti­cal intel­li­gence— where troops would land, where they would bivouac, where the hos­pi­tals and the police sta­tions were sit­u­at­ed. Army offi­cers met with the police offi­cials to see where trou­ble might occur. They talked with police offi­cers down to the precinct lev­el to spot gun shops and liquor stores that might be tar­gets for riot­ers.
...

Anoth­er goal of Coun­ter­in­tel­li­gence Analy­sis Detach­ment was pre­dict­ing when and where a civ­il dis­tur­bance might break out. The MLK was assas­si­nat­ed and protests broke out in 100 cities, mak­ing clear that pre­dict­ing civ­il dis­tur­bances when and where civ­il dis­tur­bances break out might not be fea­si­ble:

...
Pos­si­ble Agi­ta­tors

In addi­tion, Army offi­cers slid into the polit­i­cal sphere by ask­ing the police for the names and pic­tures of pos­si­ble riot agi­ta­tors. They also asked the police for the names and pic­tures of peo­ple who might be will­ing to help calm a crowd.

That infor­ma­tion, along with oth­er mate­r­i­al from the F.B.I. and the Secret Ser­vice, was fed back to Wash­ing­ton, where it went into the com­pendi­um” com­piled by the Coun­ter­in­tel­li­gence Analy­sis Detach­ment. “The com­pendi­um” was a two-vol­ume ency­clo­pe­dia that con­tained pic­tures and data, includ­ing the polit­i­cal beliefs, on peo­ple who might either foment or help stop a civ­il dis­tur­bance. The coun­ter­in­tel­li­gence group was also charged by Mr. Mc Gif­fert with try­ing to pre­dict when and where a civ­il dis­tur­bance might break out.

But the assas­si­na­tion of the Rev. Dr. Mar­tin Luther King Jr., in April, 1968, put an end to that idea. The riot­ing that occurred in 100 cities after his assas­si­na­tion showed that the site of a civ­il dis­tur­bance could not be pre­dict­ed.
...

In 1968, Under Sec­re­tary of the Army, David E. McGif­fert order the Army to be pre­pared to send 10,000 troops on short notice to 25 Amer­i­can cities:

...
High-Lev­el Review

Although the Army was bet­ter pre­pared to han­dle the dis­or­ders in Wash­ing­ton, Bal­ti­more and Chica­go than it had been dur­ing ear­li­er riots in Newark and Detroit, the need for Fed­er­al troops and the nation­wide ten­sion stim­u­lat­ed an oth­er high-lev­el review. At meet­ings in the Pen­ta­gon on April 12 and in the White House on April 15, 1968, Mr. McGif­fert pro­posed that Army intel­li­gence con­cen­trate on civ­il dis­tur­bance warn­ings.

Out of those meet­ings also came a require­ment that the Army be pre­pared to send 10,000 troops on short notice to any one of 25 cities. That num­ber was lat­er reduced to about 10 cities where the Nation­al Guard and the local police were con­sid­ered unable to han­dle things on their own.

Through the sum­mer of 1968, Army intel­li­gence oper­a­tions inten­si­fied. The Army put into effect its civ­il dis­tur­bance in for­ma­tion plan on May 2, giv­ing its agents more col­lec­tion require­ments.

They were told to report on every­thing that bore there motest con­nec­tion to civ­il dis­tur­bances. Maj. Gen. William H. Blake­field, the intel­li­gence com­man­der, told his sub­or­di­nates to “beat the Asso­ci­at­ed Press” in their report­ing.

Gen­er­al Yarbor­ough set up a task force in the Coun­ter­in­tel­li­gence Analy­sis Detach­ment to study infor­ma­tion about the 1968 poor peo­ple’s cam­paign and Res­ur­rec­tion City in Wash­ing­ton, which were close­ly scru­ti­nized by mil­i­tary intel­li­gence agents.

The intel­li­gence com­mand start­ed dis­trib­ut­ing its “black list,” which includ­ed names, pic­tures, per­son­al data and polit­i­cal char­ac­ter­i­za­tions, such as “rad­i­cal” or “mil­i­tant,” of poten­tial trou­ble­mak­ers. The “black­list” went to law enforce­ment agen­cies at all lev­els, as well as Army com­man­ders and mil­i­tary intel­li­gence groups.
...

Lat­er that year, RFL was assas­si­nat­ed and Con­gress passed a res­o­lu­tion giv­ing the Secret Ser­vice the author­i­ty to draw on the Army to pro­tect nation­al polit­i­cal can­di­dates. ON June 8, 1968, Paul Nitze, the Deputy Sec­re­tary of Defense, signed an order that gave for­mal instruc­tions to pro­vide the Pen­ta­gon with all essen­tial intel­li­gence data on civ­il dis­tur­bances. This led to the cre­ation of com­put­er data­bas­es of civ­il dis­tur­bances and infor­ma­tion on indi­vid­u­als are inter­est

...
In June, 1968, Sen­a­tor Robert F. Kennedy was assas­si­nat­ed and Con­gress passed a res­o­lu­tion giv­ing the Secret Ser­vice the author­i­ty to draw on the Army and oth­er Fed­er­al agen­cies for help in pro­tect­ing nation­al polit­i­cal can­di­dates.

Paul Nitze, the Deputy Sec­re­tary of Defense, signed an order on June 8 that was the most explic­it direc­tive until then on the Army’s intel­li­gence gath­er­ing pro­ce­dures. The order gave for­mal instruc­tions to pro­vide to the Pen­ta­gon all of the essen­tial intel­li­gence data on civ­il dis­tur­bances.

The intel­li­gence com­mand at Fort Holabird began using com­put­ers to store infor­ma­tion on civ­il dis­tur­bances. One data bank con­tained a file on inci­dents, a sec­ond a bio­graph­i­cal file on sol­diers who were con­sid­ered pos­si­ble dis­senters.

A sim­i­lar data bank was opened at the Con­ti­nen­tal Army Com­mand head­quar­ters at Fort Mon­roe, Va., for a pro­gram called Rita, for Resis­tance in the Army. Still anoth­er data bank was at III Corps head quar­ters at Fort Hood, Tex. This data bank con­cen­trat­ed on civ­il dis­tur­bance infor­ma­tion because two Army divi­sions at Fort Hood had antiri­ot respon sibil­i­ties.
...

Also in June of 1968, the Direc­torate for Civ­il Dis­tur­bance Plan­ning and Oper­a­tions was set up at the Pen­ta­gon. This becamse known as the “domes­tic war room”. By the end of 1968, this whole oper­a­tion gave Army intel­li­gence the infor­ma­tion it need­ed to pre­dict how many pro­tes­tors were going to show up for a planned counter-demon­stra­tion for Nixon’s inau­gu­ra­tion and what they were plan­ning on doing:

...
‘Domes­tic War Room’

The Direc­torate for Civ­il Dis­tur­bance Plan­ning and Oper­a­tions was set up in June, 1968, in what came to be known as the “domes­tic war room” in the base­ment of the Pen­ta­gon. This group was respon­si­ble for order­ing air­lifts, troop deploy­ment and logis­tics in a civ­il dis­or­der and became a major con­sumer of Conus Intel’s in for­ma­tion.

...

By the end of 1968, the Army intel­li­gence oper­a­tion was mov­ing at top speed. When dis­senters planned their counter-inau­gur­al demon­stra­tions in Wash­ing­ton in Jan­u­ary, 1969, the Army knew how many pro­test­ers would show up and what they planned to do.

Through­out 1969, Army intel­li­gence turned out an aver­age of 1,200 spot reports each month on inci­dents around the nation. By that time, there were exten­sive inci­dent and per­son­al­i­ty files in every mil­i­tary intel­li­gence field, region­al and group head­quar­ters, plus the com­put­er banks at Fort Holabird, Fort Mon­roe and Fort Hood. In addi­tion, the counter-intel­li­gence detach­men­t’s 120,000 pages of micro­film con­tained about 5,000 pages on civil­ians.
...

But it was also around this time, right when this vast sur­veil­lance bureau­cra­cy was set and up deliv­er­ing results, that Gen­er­al Yarbor­ough was replaced by Maj. Gen. Joseph A. McChris­t­ian. After being briefed on the oper­a­tion, McChris­t­ian ordered that ways be found to cut back on this vast domes­tic sur­veil­lance oper­a­tion. But the “domes­tic war room” and Jus­tice Depart­ment pushed back, argu­ing they need­ed this infor­ma­tion:

...
Laws of Physics

Even as the Army intel­li­gence oper­a­tion was speed­ing along, how­ev­er, some efforts were begin­ning to be made to slow it down. But stop­ping it proved dif­fi­cult. Bureau­cra­cies seem to fol­low the laws of physics—a bureau­cra­cy at rest tends to stay at rest; a bureau­cra­cy in motion tends to remain in motion.

Gen­er­al Yarbor­ough, who had start­ed the Army’s domes­tic intel­li­gence oper­a­tions, was replaced in August, 1968, by Maj. Gen. Joseph A. McChris­t­ian, a for­mer head of all mil­i­tary intel­li­gence in Viet­nam.

In tak­ing over his new as sign­ment, Gen­er­al McChris­t­ian was briefed on Con­ti­nen­tal Unit­ed States Intel­li­gence, and imme­di­ate­ly asked his sub­or­di­nates to find ways to cut it back. He was pri­mar­i­ly con­cerned with the time it was tak­ing away from oth­er tasks in mil­i­tary intel­li­gence.

But the gen­er­al ran into resis­tance from the “domes­tic war room” and oth­er Gov­ern­ment agen­cies, par­tic­u­lar­ly the Jus­tice Depart­ment, that said they need­ed the infor­ma­tion com­ing from the intel­li­gence oper­a­tion.
...

Final­ly, in Feb­ru­ary of 1969, Under Sec­re­tary McGif­fert won­dered whether the Army might be exceed­ing its author­i­ty and ordered that covert oper­a­tions end:

...
Request for Film

The Under Sec­re­tary of the Army, Mr. McGif­fert, start­ed to won­der, how­ev­er, about the pro­pri­ety of the oper­a­tion in Octo­ber, when he dis­cov­ered that agents had filmed a demon­stra­tion dur­ing the Demo­c­ra­t­ic Nation­al Con­ven­tion in Chica­go. That came to his atten­tion when the Jus­tice Depart­ment asked him for the video tape for pos­si­ble use in the Chica­go Sev­en tri­al.

In a mem­o­ran­dum dat­ed Feb. 5, 1969, short­ly before he left the Gov­ern­ment after the Nixon Admin­is­tra­tion took over, Mr. McGif­fert said that mil­i­tary intel­li­gence might be in dan­ger of exceed­ing its author­i­ty and that hence­forth no covert oper­a­tions would be con­duct­ed.

Short­ly after, the intel­li­gence com­mand stopped dis­trib­ut­ing its “black­list,” but kept it up to date until the end of 1969, when it was ordered with­drawn. Because so many copies had been sent out, the Pen­ta­gon could not be sure that they were all returned for destruc­tion.
...

But around the same time McGif­fert ordered the end of the covert oper­a­tions, the Army gen­er­al coun­sel explored with the Deputy Attor­ney Gen­er­al Richard G. Klein­di­enst whether or not the Jus­tice Depart­ment could take over these intel­li­gence gath­er­ing oper­a­tions. Klein­di­enst assert­ed that the Jus­tice Depart­ment lacked the man­pow­er:

...
Agree­ment Sought

About the same time, Mr. Jor­dan, the Army gen­er­al coun­sel, began explor­ing with Deputy Attor­ney Gen­er­al Richard G. Klein­di­enst the pos­si­bil­i­ty of hav­ing the Jus­tice Depart­ment take over intel­li­gence gath­er­ing on civ­il dis­tur­bances. But at a meet­ing on April 1, he was unable to obtain an agree­ment. Mr. Klein­di­enst con­tend­ed that his depart­ment lacked the man­pow­er to do the job.

Nev­er­the­less, Gen­er­al McChris­t­ian ordered some restric­tions on his own author­i­ty, instruct­ing that more time be put on secu­ri­ty clear­ances and oth­er tasks of pro­tect­ing Army instal­la­tions.
...

And it was­n’t just pro­tes­tors and dis­si­dents who were tar­get­ed for sur­veil­lance. Sen­a­tor Sam J. Ervin Jr., Demo­c­rat of North Car­oli­na, charged that CONUS Intel was also spy­ing on politi­cians:

...
The Army now autho­rizes only lim­it­ed intel­li­gence gath­er­ing on inci­dents that might lead to a Pres­i­den­tial call for Fed­er­al troops. But atten­tion was sharply focused last month on the Army oper­a­tion when Sen­a­tor Sam J. Ervin Jr., Demo­c­rat of North Car­oli­na, con­tend­ed that promi­nent polit­i­cal fig­ures in Illi­nois had been under mil­i­tary sur­veil­lance since 1968.

Sen­a­tor Ervin is skep­ti­cal of the Army’s announce­ment about halt­ing the spy­ing and has sched­uled hear­ings by the Sub­com­mit­tee on Con­sti­tu­tion­al Rights, of which he is chair­man, to begin Feb. 23. He said the Army “must dis­close in full what hap­pened and why it hap­pened and what has been done to insure that it will nev­er hap­pen again.”

...

Things were qui­et then until last month, when Sen­a­tor Ervin opened the issue up again by charg­ing that Con­ti­nen­tal Unit­ed States intel­li­gence had spied on Sen­a­tor Adlai E. Steven­son 3d, Rep­re­sen­ta­tive Abn­er Mik­va, and for­mer Gov. Otto Kern­er, all of Illi­nois.

Sen­a­tor Erv­in’s alle­ga­tions were based on state­ments by John M. O’Brien, a for­mer staff sergeant who had served with the 113th Group in Chica­go. He made them again under oath in a suit brought by the A.C.L.U. in Chica­go. But Sec­re­tary Resor cat­e­gor­i­cal­ly denied the charges.
...

And that’s what we learned about this oper­a­tion back in 1971. So were the lessons of this expe­ri­ence actu­al­ly learned by the Amer­i­can peo­ple? We’ll prob­a­bly find out as we see this ‘Trump unleashed’ peri­od of Trump’s pres­i­den­cy unfold. But it’s pret­ty clear that all of the pieces are in place for a major domes­tic oper­a­tion that uti­lizes the full pow­er of the state to attack the per­ceived polit­i­cal ene­mies of the White House. Trump him­self has now made this clear.

Also note how the assas­si­na­tions of MLK and RFK played into the jus­ti­fi­ca­tion for this domes­tic mil­i­tary intel­li­gence oper­a­tion. The mass riots that broke out in cities across the coun­try only fueled the call for the capa­bil­i­ty of send­ing thou­sands of troops into a large num­ber of cities in short order simul­ta­ne­ous­ly and RFK’s assas­si­na­tion result­ed in Con­gress­ing allow­ing the Secret Ser­vice to call in fed­er­al troops to pro­tect can­di­dates. And yet both the MLK and RFK assas­si­na­tions had gov­ern­ment fin­ger­prints all over them. See AFA #46 for much more on the gov­ern­ment role in the assas­si­na­tion of MLK (part 2 includes ref­er­ences to Yarbor­ough and the domes­tic mil­i­tary intel­li­gence gath­er­ing going on at this time) and FTR#789 for more on RFK’s assas­si­na­tion and how the US’s pro­gres­sive lead­er­ship was being sys­tem­at­i­cal­ly killed out dur­ing this peri­od. It’s a huge and dark aspect of the sto­ry of CONUS Intel: it was a mil­i­tary intel­li­gence oper­a­tion that did­n’t just involve gath­er­ing mas­sive amounts of data on domes­tic dis­si­dents. It also involved plan­ning for fed­er­al troops to move into cities and took place dur­ing a peri­od when Amer­i­can’s left-wing lead­ers were get­ting killed off by right-wing forces oper­at­ing with­in and out­side the gov­ern­ment (i.e. the real ‘Deep State’). So now that Pres­i­dent Trump has made it clear that he feels embold­ened to do pret­ty much what­ev­er he wants to do against his oppo­nents, now is prob­a­bly a good time for Amer­i­can to revis­it Amer­i­ca’s long his­to­ry of cod­dling fas­cists while overt­ly and covert­ly using the full pow­er of the nation­al secu­ri­ty state against for domes­tic polit­i­cal agen­das. Domes­ti­cal­ly polit­i­cal­ly agen­das that were vir­tu­al­ly always tar­get­ing pro­gres­sives.

Here’s a pair of sto­ries that high­light one of the Big Data areas of infor­ma­tion Palan­tir is giv­en access to by the US gov­ern­ment: mas­sive vol­umes of IRS infor­ma­tion.

First, here’s an arti­cle from Decem­ber 2018 about how the IRS has turned to Palan­tir’s AI to find signs of tax fraud. The IRS had signed a $99 mil­lion sev­en year con­tract with Palan­tir that Sep­tem­ber. The con­tract let’s the IRS search for tax cheats using Palan­tir’s soft­ware by min­ing tax returns, bank reports, prop­er­ty records and even social media posts. As the arti­cle notes, part of the motive for rely­ing on Palan­tir for this work is because the IRS has seen its staff shrink so much in recent years, with over $1 bil­lion in cuts to the IRS bud­get since 2010. The Crim­i­nal Inves­ti­ga­tions divi­sion has lost around 150 agents per year as a result of these cuts. These IRS bud­get cuts, of course, are the work of the Repub­li­cans. As a result, AI and machine learn­ing approach­es to find­ing crim­i­nal activ­i­ty are now seen as nec­es­sary for the IRS to do its job with few­er staff and resources. It sounds like the Palan­tir sys­tems have access to the IRS’s Com­pli­ance Data Ware­house, which has 40 data sets on tax­pay­ers stretch­ing back more than 30 years.

Now, the idea of using AI and machine learn­ing in the IRS’s crim­i­nal divi­sion seems like a very rea­son­able approach in gen­er­al. But this isn’t the IRS imple­ment­ing these approach­es. This is the IRS hand­ing over vast vol­umes of data to Palan­tir and using Palan­tir’s tools to do the analy­sis. Which implies Palan­tir has access to these IRS data­bas­es and, in turn, implic­it con­trol over which poten­tial cas­es get flagged for review by the IRS. So the IRS’s bud­get and staff get slashed and the result is the effec­tive pri­va­ti­za­tion of the IRS’s crime detec­tion capa­bil­i­ties. And the com­pa­ny that is pro­vid­ing these crime detec­tion capa­bil­i­ties is owned by Peter Thiel, a top Repub­li­can donor and one of the biggest anti-tax fas­cists in the world:

Law 360

AI Help­ing IRS Detect Tax Crimes With Few­er Resources

By Vidya Kau­ri
Decem­ber 5, 2018, 8:58 PM EST

Law360 (Decem­ber 5, 2018, 8:58 PM EST) — Under the weight of bud­get cuts and decreas­ing staff, the Inter­nal Rev­enue Ser­vice may be find­ing its salve in arti­fi­cial intel­li­gence tech­nol­o­gy to detect crim­i­nal tax activ­i­ties more effi­cient­ly.

Research and inves­tiga­tive tasks that might have tak­en weeks or months in the past may be accom­plished with­in min­utes now thanks to machine learn­ing tech­nol­o­gy the IRS is deploy­ing to detect tax­pay­er non­com­pli­ance, iden­ti­ty theft, mon­ey laun­der­ing and oth­er crim­i­nal activ­i­ties. The tools are also able to detect instances of fraud or poten­tial­ly hid­den tax­pay­er assets that rev­enue agents may not be able to dis­cov­er man­u­al­ly, IRS offi­cials said dur­ing a web­cast host­ed Wednes­day by the Amer­i­can Bar Asso­ci­a­tion.

The speed of pro­cess­ing data and con­duct­ing inves­ti­ga­tions has now picked up to the extent that agents have been able to catch per­pe­tra­tors “in the act of get­ting mon­ey from a bank,” accord­ing to Todd Egaas, direc­tor of tech­nol­o­gy, oper­a­tions and inves­tiga­tive ser­vices in the IRS’ Crim­i­nal Inves­ti­ga­tions office.

“We’ve been run­ning thin on peo­ple late­ly and rich on data,” Egaas said. “And so what we’ve been work­ing on — and this is where we think data can help us — is how do we make the most use out of our peo­ple?”

The infor­ma­tion that Egaas and his col­league Ben­jamin Hern­don, the IRS’ chief ana­lyt­ics offi­cer, shared is the first major glimpse of how the rev­enue agency is using advanced tech­nol­o­gy since it signed a sev­en-year, $99 mil­lion deal with Palan­tir Tech­nolo­gies in Sep­tem­ber to sniff out tax cheats by min­ing data from tax returns, bank reports, prop­er­ty records and even social media posts.

Hern­don, who is also the direc­tor of the IRS’ Research, Applied Ana­lyt­ics and Sta­tis­tics divi­sion, said that the agency is using machine learn­ing algo­rithms and arti­fi­cial intel­li­gence to iden­ti­fy pat­terns in graphs where non­com­pli­ance might be present. Such “graph data tech­niques” have proved par­tic­u­lar­ly help­ful in com­bat­ing iden­ti­ty thieves fraud­u­lent­ly apply­ing for tax refunds, he said.

“One of the things that we have to do to catch iden­ti­ty thieves is to under­stand the pat­tern of behav­ior that they engage in and how we can catch it before we’re already vic­tims of iden­ti­ty theft,” he said.

The IRS is also using nat­ur­al lan­guage pro­cess­ing — tech­nol­o­gy that can enable a com­put­er to read — to trans­late fil­ings in for­eign lan­guages, par­tic­u­lar­ly with­in its Large Busi­ness and Inter­na­tion­al divi­sion. Although the tech­nol­o­gy has more pow­er­ful uses to ana­lyze the text of these doc­u­ments, instead of just con­vert­ing them into some­thing under­stand­able, Hern­don said the agency is not yet ready to use it for ana­lyt­ics.

“We have to make sure the trans­la­tion tech­nol­o­gy works first,” he said.

The IRS is under sig­nif­i­cant pres­sure to be smarter in choos­ing cas­es that are more like­ly to con­clude well. The agency has faced con­sec­u­tive bud­get cuts since 2010, amount­ing to a loss of more than $1 bil­lion. The loss is acute­ly felt with­in the Crim­i­nal Inves­ti­ga­tions divi­sion, which los­es about 150 agents a year to attri­tion, accord­ing to Egaas.

With about 2,100 agents remain­ing, it is espe­cial­ly cru­cial for the divi­sion to use advanced machine lan­guage algo­rithms, nat­ur­al lan­guage pro­cess­ing and graph ana­lyt­ics.

While Palantir’s tech­nol­o­gy to con­nect rela­tion­ships between dif­fer­ent enti­ties across mul­ti­ple data sources has been used only with­in the CI office so far, the RAAS divi­sion has been cre­at­ing more inex­pen­sive tech­nolo­gies using open source graph data with­in a pro­pri­etary envi­ron­ment. These tools help the IRS iden­ti­fy non­com­pli­ance or help explain anom­alies in indi­vid­ual fil­ings.

For exam­ple, a fraud tech­ni­cal advi­sor was able to iden­ti­fy 296 sus­pect­ed iden­ti­ty theft returns claim­ing $1.3 mil­lion using such tech­nol­o­gy, accord­ing to Hern­don. Pre­vi­ous­ly, 84 per­cent of these had not been iden­ti­fied.

The infra­struc­ture that sup­ports RAAS’ explo­ration and pro­to­typ­ing activ­i­ties is sup­port­ed by a Com­pli­ance Data Ware­house, which was orig­i­nal­ly built in the 1990s to col­lect data from mul­ti­ple data­bas­es and begin using for­ward-look­ing data analy­sis to pre­vent fraud. The CDW has about 40 data sets on tax­pay­ers stretch­ing back more than 30 years.

“This is a large data set with very rich meta­da­ta over it, which is crit­i­cal to mak­ing it use­ful. I think the cur­rent size is some­thing like six petabytes, which makes it a pret­ty rich resource,” Hern­don said.

Nat­ur­al lan­guage pro­cess­ing and text ana­lyt­ics are also being used in appeal cas­es to try to under­stand why a par­tic­u­lar case was lost or won, and then to pre­dict the like­li­hood of win­ning, and on what grounds, in oth­er sim­i­lar cas­es.

The IRS is also using algo­rithms sim­i­lar to those used by online busi­ness­es that can make rec­om­men­da­tions of items to buy based on what a con­sumer has pre­vi­ous­ly bought or searched for. In the IRS’ case, these algo­rithms are used to fill in gaps in data or indi­vid­ual val­ues on tax returns based on a pat­tern of sur­round­ing data. This can be done in real time, once a machine is ade­quate­ly trained, to iden­ti­fy anom­alies and pat­terns as they come in, Hern­don said.

This rec­om­men­da­tion machine can also be work­ing in the back­ground look­ing for records that might be of inter­est while an agent looks at a case, and the agent can enhance the machine’s learn­ing by let­ting it know if the records dug up are tru­ly rel­e­vant or not, Egaas said.

“The rec­om­men­da­tion engine for us helps us avoid blind spots,” Egaas said. “We are a small agency tasked with enforc­ing tax law across 250 mil­lion Amer­i­cans. It’s easy to get blind spots and tech­nol­o­gy is help­ing address that for us.”

In June, the IRS issued a request for infor­ma­tion seek­ing com­ments on AI, machine learn­ing, cus­tomiz­able user inter­faces and cloud com­put­ing. The agency said that it was look­ing for ways to pro­vide con­text for alerts or cas­es used for inves­ti­ga­tion, iden­ti­fy pre­vi­ous­ly unknown threats and sup­port stream­ing data sources to pro­vide near real-time assess­ment with­in 48 hours.

As the agency expands its tech­no­log­i­cal capa­bil­i­ties, the IRS offi­cials stressed that tax­pay­er pri­va­cy remains para­mount and that any ven­dors it con­tracts with have to go through the same secu­ri­ty and com­pli­ance checks that IRS staff must go through. Even data that is main­tained in cloud sys­tems is under the full con­trol of the IRS, they said.

...

———-

“AI Help­ing IRS Detect Tax Crimes With Few­er Resources” by Vidya Kau­ri; Law 360; 12/05/2018

“The infor­ma­tion that Egaas and his col­league Ben­jamin Hern­don, the IRS’ chief ana­lyt­ics offi­cer, shared is the first major glimpse of how the rev­enue agency is using advanced tech­nol­o­gy since it signed a sev­en-year, $99 mil­lion deal with Palan­tir Tech­nolo­gies in Sep­tem­ber to sniff out tax cheats by min­ing data from tax returns, bank reports, prop­er­ty records and even social media posts.”

After $1 bil­lion in IRS cuts over the past eight years, the IRS signs a 7 year $99 mil­lion deal with Palan­tir to help make up for the lost man­pow­er. It’s a pret­ty nice deal with Palan­tir, which now has access to more than three decades of infor­ma­tion from the IRS’s Com­pli­ance Data Ware­house:

...
The IRS is under sig­nif­i­cant pres­sure to be smarter in choos­ing cas­es that are more like­ly to con­clude well. The agency has faced con­sec­u­tive bud­get cuts since 2010, amount­ing to a loss of more than $1 bil­lion. The loss is acute­ly felt with­in the Crim­i­nal Inves­ti­ga­tions divi­sion, which los­es about 150 agents a year to attri­tion, accord­ing to Egaas.

With about 2,100 agents remain­ing, it is espe­cial­ly cru­cial for the divi­sion to use advanced machine lan­guage algo­rithms, nat­ur­al lan­guage pro­cess­ing and graph ana­lyt­ics.

While Palantir’s tech­nol­o­gy to con­nect rela­tion­ships between dif­fer­ent enti­ties across mul­ti­ple data sources has been used only with­in the CI office so far, the RAAS divi­sion has been cre­at­ing more inex­pen­sive tech­nolo­gies using open source graph data with­in a pro­pri­etary envi­ron­ment. These tools help the IRS iden­ti­fy non­com­pli­ance or help explain anom­alies in indi­vid­ual fil­ings.

...

The infra­struc­ture that sup­ports RAAS’ explo­ration and pro­to­typ­ing activ­i­ties is sup­port­ed by a Com­pli­ance Data Ware­house, which was orig­i­nal­ly built in the 1990s to col­lect data from mul­ti­ple data­bas­es and begin using for­ward-look­ing data analy­sis to pre­vent fraud. The CDW has about 40 data sets on tax­pay­ers stretch­ing back more than 30 years.

“This is a large data set with very rich meta­da­ta over it, which is crit­i­cal to mak­ing it use­ful. I think the cur­rent size is some­thing like six petabytes, which makes it a pret­ty rich resource,” Hern­don said.
...

And we’re assured that any ven­dors the IRS con­tracts with to car­ry out its tasks will have to go through the same secu­ri­ty and com­pli­ance checks that IRS staff go through because pri­va­cy is para­mount. So don’t wor­ry about giv­ing even more infor­ma­tion to Palan­tir because its employ­ees giv­en access to this data have to go through secu­ri­ty checks. That’s the lev­el of assur­ance we’re get­ting about hand­ing over this vast amount of finan­cial data to a com­pa­ny run by a lib­er­tar­i­an fas­cist:

...
As the agency expands its tech­no­log­i­cal capa­bil­i­ties, the IRS offi­cials stressed that tax­pay­er pri­va­cy remains para­mount and that any ven­dors it con­tracts with have to go through the same secu­ri­ty and com­pli­ance checks that IRS staff must go through. Even data that is main­tained in cloud sys­tems is under the full con­trol of the IRS, they said.
...

This is also a good time to recall the sto­ry about JP Mor­gan hir­ing Palan­tir to pro­vide AI over­sight of JP Mor­gan’s employ­ees. It turns out the JP Mor­gan secu­ri­ty offi­cer who was giv­en access to Palan­tir’s obser­va­tion sys­tems, Peter Cav­ic­chia, ‘went rogue’ and start­ed spy­ing on peo­ple all over the com­pa­ny, includ­ing the exec­u­tives. Cav­ic­chia had a team of Palan­tir employ­ees work­ing for him and unprece­dent­ed access to the bank’s inter­nal infor­ma­tion, like emails, and the Palan­tir sys­tem had no real lim­its. Cav­ic­chia went wild spy­ing on peo­ple at the bank, result­ing in JP Mor­gan cur­tail­ing its use of Palan­tir’s sys­tems. That’s the kind of com­pa­ny that’s being trust­ed with these data­bas­es of US tax records. And keep in mind that there’s noth­ing stop­ping Palan­tir from com­bin­ing the infor­ma­tion it gets from the IRS with the finan­cial infor­ma­tion its get­ting from the banks too. It’s lit­er­al­ly posi­tioned to become the lead­ing Big Data pri­vate repos­i­to­ry of sen­si­tive infor­ma­tion and its run by a Trump-lov­ing fas­cist.

Now here’s an exam­ple of, iron­i­cal­ly, an IRS work­er who was just sen­tenced to five years pro­ba­tion for leak­ing an IRS “sus­pi­cious activ­i­ty report”. The IRS analy­sis, John Fry, is charged with pulling a “sus­pi­cious activ­i­ty report” report relat­ed to Pres­i­dent Trump’s per­son­al attor­ney, Michael Cohen. Fry grabbed the report from a con­fi­den­tial law enforce­ment data­base and leaked it to Stormy Daniel’s attor­ney, Michael Ave­nat­ti, in May of 2018. Fry grabbed the reports from the Palan­tir data­base used by the IRS Crim­i­nal Inves­ti­ga­tion divi­sion. It’s an exam­ple of the kind of poten­tial­ly polit­i­cal pow­er­ful infor­ma­tion Palan­tir was giv­en access to with its IRS con­tract:

Cour­t­house News

IRS Work­er Who Leaked Cohen Docs Sen­tenced to Five Years Pro­ba­tion

NICHOLAS IOVINO
Jan­u­ary 17, 2020

SAN FRANCISCO (CN) – An IRS employ­ee accused of ille­gal­ly leak­ing for­mer Trump attor­ney Michael Cohen’s bank infor­ma­tion to Stormy Daniels’ lawyer was sen­tenced to five years pro­ba­tion Fri­day after plead­ing guilty to one count of dis­clos­ing unau­tho­rized doc­u­ments.

Pros­e­cu­tors with the Depart­ment of Trea­sury say John Fry, an IRS ana­lyst, pulled “sus­pi­cious activ­i­ty reports” relat­ed to Cohen’s accounts from con­fi­den­tial law enforce­ment data­bas­es and leaked the infor­ma­tion to Stormy Daniels’ attor­ney, Michael Ave­nat­ti, in May 2018.

On May 8, Ave­nat­ti post­ed per­son­al doc­u­ments relat­ed to Cohen and Cohen’s shell com­pa­ny, Essen­tial Con­sul­tants LLC, on Twit­ter. From there, the Wash­ing­ton Post picked up the infor­ma­tion, fol­lowed by an inves­tiga­tive sto­ry by the New York­er.

Essen­tial Con­sul­tants is the com­pa­ny Cohen used to pay porn actor Stormy Daniels to keep qui­et about her alleged affair with Pres­i­dent Don­ald Trump.

The May 8 Wash­ing­ton Post sto­ry not­ed that Ave­nat­ti refused to reveal where he received his doc­u­ments, say­ing, “The source or sources of our infor­ma­tion is our work prod­uct, and nobody’s busi­ness … They can inves­ti­gate all they want, but what they should be doing is releas­ing to the Amer­i­can pub­lic the three Sus­pi­cious Activ­i­ty Reports filed on Michael Cohen’s account.”

Fry has worked for the IRS since 2008 and was work­ing in the agency’s San Fran­cis­co office as of Feb­ru­ary last year. As an IRS ana­lyst, he had access to var­i­ous law enforce­ment data­bas­es, includ­ing the Palan­tir data­base used by the IRS Crim­i­nal Inves­ti­ga­tion divi­sion to col­lect inves­tiga­tive data from mul­ti­ple sources, accord­ing to a crim­i­nal com­plaint filed in Feb­ru­ary 2019.

Pros­e­cu­tors claimed Fry logged into the Palan­tir data­base at 2:54 p.m. on May 4, 2018, and down­loaded five sus­pi­cious activ­i­ty reports relat­ed Michael Cohen and his shell com­pa­ny. One of the reports detailed deposits of three checks in the amounts of $505,000, $250,000 and $250,000 from the shell company’s First Repub­lic Bank account.

Imme­di­ate­ly after down­load­ing the files, Fry placed two phone calls to Daniels’ lawyer Ave­nat­ti. He lat­er had a phone con­ver­sa­tion with and exchanged texts with an unnamed reporter on the What­sApp mes­sag­ing app, accord­ing to the com­plaint.

Pros­e­cu­tors said Fry also searched for and attempt­ed to retrieve oth­er unau­tho­rized reports in a sep­a­rate crim­i­nal data­base, but because those reports were restrict­ed, he could not access them.

Accord­ing to the com­plaint, Fry ver­bal­ly con­fessed to leak­ing the reports when con­front­ed by two spe­cial agents at the IRS office in San Fran­cis­co on Nov. 26, 2018.

Fry was charged with one count of unau­tho­rized dis­clo­sure of sus­pi­cious activ­i­ty reports, two counts of mis­use of a com­put­er and one count of unau­tho­rized use of a social secu­ri­ty num­ber. He faced a max­i­mum 20 years in prison and $1 mil­lion fine.

Pros­e­cu­tors dropped three of the charges after Fry agreed to plead guilty to one count of unau­tho­rized use of sus­pi­cious activ­i­ty reports in August 2019.

Fry will pay a $5,000 fine and be sub­ject to a five-year pro­ba­tion term with stan­dard pro­ba­tion con­di­tions imposed, includ­ing a require­ment that he not leave the North­ern Dis­trict of California’s bound­aries with­out advanced per­mis­sion from his pro­ba­tion offi­cer.

...

———–

“IRS Work­er Who Leaked Cohen Docs Sen­tenced to Five Years Pro­ba­tion” by NICHOLAS IOVINO; Cour­t­house News; 01/17/2020

“Fry has worked for the IRS since 2008 and was work­ing in the agency’s San Fran­cis­co office as of Feb­ru­ary last year. As an IRS ana­lyst, he had access to var­i­ous law enforce­ment data­bas­es, includ­ing the Palan­tir data­base used by the IRS Crim­i­nal Inves­ti­ga­tion divi­sion to col­lect inves­tiga­tive data from mul­ti­ple sources, accord­ing to a crim­i­nal com­plaint filed in Feb­ru­ary 2019.”

Yep, as an IRS ana­lyst, Fry had access to var­i­ous law enforce­ment data­bas­es, includ­ing the Palan­tir data­base used by the IRS Crim­i­nal Inves­ti­ga­tion divi­sion to col­lect inves­tiga­tive data from mul­ti­ple sources. IRS ana­lysts have access to those data­bas­es and now Palan­tir employ­ees have access too thanks to these kinds of con­tracts with the IRS. And Fry tried to access even more reports in a sep­a­rate crim­i­nal data­base but those reports were restrict­ed. It rais­es the ques­tion of whether or not that sep­a­rate restrict­ed data­base was one of the data­bas­es main­tained by Palan­tir or not. Because if it was main­tained by Palan­tir we should keep in mind that Palan­tir’s engi­neers pre­sum­ably have access to those restrict­ed files even if IRS agents like Fry don’t have access. It’s one of the caveats with the assur­ances we get that the employ­ees for ven­dors like Palan­tir who are giv­en access to these data­bas­es are going to go through secu­ri­ty checks like gov­ern­ment employ­ees. Those Palan­tir employ­ees might effec­tive­ly access to ALL of the infor­ma­tion that their gov­ern­ment employ­ee coun­ter­parts can’t nec­es­sar­i­ly access so if a Palan­tir employ­ee ‘goes rogue’ the dam­age they could do is prob­a­bly far greater than an IRS or oth­er gov­ern­ment employ­ee going rogue:

...
Pros­e­cu­tors claimed Fry logged into the Palan­tir data­base at 2:54 p.m. on May 4, 2018, and down­loaded five sus­pi­cious activ­i­ty reports relat­ed Michael Cohen and his shell com­pa­ny. One of the reports detailed deposits of three checks in the amounts of $505,000, $250,000 and $250,000 from the shell company’s First Repub­lic Bank account.

Imme­di­ate­ly after down­load­ing the files, Fry placed two phone calls to Daniels’ lawyer Ave­nat­ti. He lat­er had a phone con­ver­sa­tion with and exchanged texts with an unnamed reporter on the What­sApp mes­sag­ing app, accord­ing to the com­plaint.

Pros­e­cu­tors said Fry also searched for and attempt­ed to retrieve oth­er unau­tho­rized reports in a sep­a­rate crim­i­nal data­base, but because those reports were restrict­ed, he could not access them.
...

Now, in this case, it was an IRS employ­ee, not a Palan­tir employ­ee, who did the leak­ing. But we have no choice in giv­ing IRS employ­ees to this infor­ma­tion. They’re sup­posed to have access to it and the risk of leaks like this is an unavoid­able risk that comes with the ter­ri­to­ry. But the risk of Palan­tir employ­ees abus­ing this kind of infor­ma­tion it’s a com­plete­ly avoid­able risk. It’s a choice to out­source these AI capa­bil­i­ties to Palan­tir. There’s no com­pelling rea­son to out­source these giant sen­si­tive data oper­a­tions. Yes, it would be more expen­sive for the IRS to devel­op­ing these kinds of AI capa­bil­i­ties on their own, but that high­er cost comes with the ben­e­fit of not hand­ing over giant data­bas­es of sen­si­tive infor­ma­tion to pri­vate com­pa­nies. At this point, Palan­tir is the AI/machine learn­ing out­sourc­ing enti­ty of choice for the US gov­ern­ment. It has the sys­tems set up to incor­po­rate new clients and teams trained to car­ry it out. And that was a choice. There’s no rea­son there could­n’t have been a gov­ern­ment agency set up to pro­vide these ser­vices to oth­er gov­ern­ment agen­cies like IRS. We could have lim­it­ed access to these vast data­bas­es to gov­ern­ment employ­ees but thanks to the reli­gion of pri­va­ti­za­tion that dom­i­nates the US gov­ern­ment Palan­tir was tapped as a Big Data/AI pri­vate out­sourc­ing enti­ty that the US gov­ern­ment could trust and now it has access to prob­a­bly more infor­ma­tion on indi­vid­ual Amer­i­cans than any oth­er sin­gle enti­ty on the plan­et. If the US gov­ern­ment set out to cre­ate a pri­va­tized ver­sion of J. Edgar Hoover’s black­mail oper­a­tion it could­n’t have done a bet­ter job than putting Peter Thiel in the posi­tion he’s in today with Palan­tir.

And that’s per­haps the biggest les­son from this to keep in mind: While grant­i­ng access to these vast troves of gov­ern­ment data­bas­es to Palan­tir employ­ees is obvi­ous­ly prob­lem­at­ic, there’s one par­tic­u­lar indi­vid­ual at Palan­tir that we need to be extra con­cerned about hav­ing access to this infor­ma­tion because he’s a fas­cist with insa­tiable per­son­al ambi­tion and appears to be amoral and more than will­ing to abuse such pow­ers if it suits his per­son­al goals. And he’s not an employ­ee. He’s the own­er.

Here’s a dis­turb­ing update on the bureau­crat­ic maneu­ver­ings involv­ing the US Under­sec­re­tary of Defense for Research and Engi­neer­ing, a lead­ing role for devel­op­ing next-gen­er­a­tion weapon sys­tems and tech­nolo­gies. First, recall how for­mer NASA admin­is­tra­tor Mike Grif­f­en was appoint­ed as act­ing Under­sec­re­tary of Defense for Research and Engi­neer­ing with an agen­da of over­haul­ing and stream­lin­ing the mil­i­tary’s defense tech­nol­o­gy pro­cure­ment process­es with the goal of facil­i­tat­ing the rapid devel­op­ment of next-gen­er­a­tion tech­nolo­gies uti­liz­ing exist­ing com­mer­cial­ly avail­able tech­nolo­gies when­ev­er pos­si­ble and reduc­ing delays caused by cost/risk assess­ments. Grif­fin was also a major advo­cate of the cre­ation of the Space Defense Agency (‘Space Force’), a favorite pet project of Pres­i­dent Trump. Also recall how Grif­fin appeared to be behind the push to end the Pen­tagon’s con­tract with the JASON group, which was part of his larg­er agen­da of min­i­miz­ing the review process for approv­ing the devel­op­ment of new plat­forms.

So Grif­fin had major visions for over­haul­ing how the US nation­al secu­ri­ty state makes deci­sions on which hi-tech projects to invest in with an eye on speed­ing the process up by rely­ing more on com­mer­cial tech­nol­o­gy and dra­mat­i­cal­ly lim­it­ing the num­ber of peo­ple involved with review­ing the pro­pos­als. And while it remains to be seen whether or not Griffin’s vision will be ful­ly real­ized, we do now know that it won’t be Grif­fin who com­pletes this vision because he just announced his res­ig­na­tion a few weeks ago, along with his deputy Lisa Porter. The news came a day after the House Armed Ser­vices Com­mit­tee rec­om­mend­ed remov­ing the Mis­sile Defense Agency from Griffin’s con­trol. So the Pen­tagon’s two top tech­nol­o­gy experts are set to be replaced:

Defense News

Pentagon’s top tech experts, Grif­fin and Porter, resign

By: Aaron Mehta and Joe Gould
June 23, 2020

WASHINGTON — The Pentagon’s top two tech­nol­o­gy experts have sub­mit­ted their res­ig­na­tions.

Mike Grif­fin, who became the Defense Department’s first under­sec­re­tary of research and engi­neer­ing in ear­ly 2018, and his deputy Lisa Porter have both sub­mit­ted their res­ig­na­tions, a defense offi­cial con­firmed to Defense News. The two will be exit­ing the build­ing July 10.

In a let­ter to R&E staff, Grif­fin and Porter wrote that “a pri­vate-sec­tor oppor­tu­ni­ty has pre­sent­ed itself to us, offer­ing an oppor­tu­ni­ty we have decid­ed to pur­sue togeth­er.”

“It has been a plea­sure lead­ing this great team over the past few years. We great­ly appre­ci­ate your hard work, dili­gence, integri­ty, and devo­tion to tech­ni­cal excel­lence and tech­ni­cal truth in fur­ther­ance of the R&E mis­sion,” the duo wrote. “We wish you all the very best.”

They become the third and fourth offi­cials to sub­mit res­ig­na­tions in the last week. On June 16, Elaine McCusker, the department’s act­ing comp­trol­ler, sub­mit­ted her res­ig­na­tion, fol­lowed two days lat­er by Kathryn Wheel­barg­er, the act­ing assis­tant defense sec­re­tary for inter­na­tion­al secu­ri­ty affairs.

...

In his role as R&E head, Grif­fin had the lead on devel­op­ing new capa­bil­i­ties for the depart­ment, such as hyper­son­ic weapons, direct­ed ener­gy and a vari­ety of space-based pro­grams. Includ­ed in his port­fo­lio were the Mis­sile Defense Agency and the Defense Advanced Research Projects Agency.

Porter, who pre­vi­ous­ly was exec­u­tive vice pres­i­dent and direc­tor of In-Q-Tel Labs, served as Griffin’s deputy from 2018 onward. Although high­ly respect­ed, she kept a low pro­file, large­ly avoid­ing media engage­ments dur­ing her time in office.

A for­mer NASA head under Pres­i­dent George W. Bush, Grif­fin entered the Depart­ment of Defense with a rep­u­ta­tion as an inno­v­a­tive thinker, but also as some­one who could be prick­ly with oth­ers. In his first pub­lic speech after tak­ing office, he infa­mous­ly bragged that he answered to no one but the sec­re­tary and deputy sec­re­tary of defense — a state­ment that ran­kled mem­bers of Con­gress.

At the DoD, Grif­fin showed a strong per­son­al­i­ty that clashed with var­i­ous ser­vice-lev­el exec­u­tives, with the most pub­lic fight com­ing with for­mer Air Force Sec­re­tary Heather Wil­son. The news emerged a day after the House Armed Ser­vices Com­mit­tee rec­om­mend­ed remov­ing the Mis­sile Defense Agency from under Griffin’s con­trol.

Mem­bers of Con­gress have expressed frus­tra­tion with Griffin’s deci­sion to can­cel the Redesigned Kill Vehi­cle. Alas­ka Repub­li­can Sen. Dan Sul­li­van said the move would mean 20 mis­sile silos at Fort Gree­ley will be emp­ty for a decade, and meant Grif­fin was out of step with the pres­i­dent.

“Heck, if you were at the Pen­ta­gon when the pres­i­dent was announc­ing the Mis­sile Defense Review, he specif­i­cal­ly men­tioned these silos at Fort Gree­ley. I’m not even sure he knows about the fact that one of his low­er-lev­el under­sec­re­taries decid­ed on his own to dig 20 holes and not put any­thing in there for at least 10 years,” Sul­li­van, a mem­ber of the Sen­ate Armed Ser­vices Com­mit­tee, told Defense News last month. “That just makes no sense. None. Zero.”

Grif­fin him­self may have hint­ed that his time at the build­ing is com­ing to a close in a May 20 speech, which he opened by not­ing that being a “pres­i­den­tial appointee, I think most of you know, is lit­er­al­ly at the plea­sure of the admin­is­tra­tion. So, we nev­er know for employ­ment is until tomor­row or next year or any­thing in between.

“But, that’s okay. You don’t you don’t take these jobs unless you under­stand that. This is my third time in that are­na,” he said.

———-

“Pentagon’s top tech experts, Grif­fin and Porter, resign” by Aaron Mehta and Joe Gould; Defense News; 06/23/2020

“In his role as R&E head, Grif­fin had the lead on devel­op­ing new capa­bil­i­ties for the depart­ment, such as hyper­son­ic weapons, direct­ed ener­gy and a vari­ety of space-based pro­grams. Includ­ed in his port­fo­lio were the Mis­sile Defense Agency and the Defense Advanced Research Projects Agency.”

As we can see, there’s going to be a new vision for the Pen­tagon’s approach to devel­op­ing new weapons, along with all the oth­er projects being devel­oped by DARPA with dual-use military/commercial appli­ca­tions. And note how Griffin’s deputy, Lisa Porter, pre­vi­ous­ly served as exec­u­tive vice pres­i­dent and direc­tor of the CIA’s pri­vate invest­ment com­pa­ny, In-Q-Tel Labs. It’s a reflec­tion of how Griffin’s vision or rely­ing more and more on read­i­ly avail­able com­mer­cial tech­nol­o­gy was like­ly going to involve more nation­al secu­ri­ty state invest­ments in the pri­vate sec­tor via com­pa­nies like In-Q-Tel:

...
Porter, who pre­vi­ous­ly was exec­u­tive vice pres­i­dent and direc­tor of In-Q-Tel Labs, served as Griffin’s deputy from 2018 onward. Although high­ly respect­ed, she kept a low pro­file, large­ly avoid­ing media engage­ments dur­ing her time in office.
...

And since this is the Trump admin­is­tra­tion that’s going to choos­ing Griffin’s replace­ment in the mid­dle of this push to cut reviews and incor­po­rate more off-the-shelf exist­ing com­mer­cial tech­nol­o­gy into the devel­op­ment of the Pen­tagon’s next-gen­er­a­tion sys­tems we have to won­der who the Trump admin­is­tra­tion is going to find, espe­cial­ly giv­en the fact that we’re months away from an elec­tion. And we just got our answer: Mike Grif­fin — who for all his faults was actu­al­ly tech­ni­cal­ly extreme­ly com­pe­tent — is going to be replaced by the White House’s chief tech­nol­o­gy offi­cer Michael Krat­sios. So is Krat­sios qual­i­fied for a posi­tion like this? Well, he’s the White House­’s chief tech­nol­o­gy office so one might assume he’s well qual­i­fied for a posi­tion like this. But as we’ll see, it turns out Krat­sios has no tech­ni­cal edu­ca­tion and his pri­mar­i­ly qual­i­fi­ca­tion is that he worked for Peter Thiel’s invest­ment com­pa­ny, Clar­i­um Cap­i­tal, and end­ed up becom­ing Thiel’s chief of staff. So the main qual­i­fi­ca­tion for next Under­sec­re­tary of Defense for Research and Engi­neer­ing is what­ev­er expe­ri­ence he acquired as an unqual­i­fied White House chief tech­nol­o­gy offi­cer. Krat­sios will con­tin­ue serv­ing as the White House­’s chief tech­nol­o­gy offi­cer. It’s the kind of sit­u­a­tion that sug­gests Krat­sios’s real qual­i­fi­ca­tions are large­ly going to be lim­it­ed to his enthu­si­asm at steer­ing more defense spend­ing towards Thiel’s com­pa­nies like Palan­tir:

Defense One

Peter Thiel’s New Man In The Defense Depart­ment

By Patrick Tuck­er Tech­nol­o­gy Edi­tor
July 13, 2020

The new head of defense research and engi­neer­ing comes from the White House with a rel­a­tive­ly light resume.

Updat­ed: 10:20 a.m.

The Pentagon’s new 33-year-old head of research and engi­neer­ing lacks a basic sci­ence degree but brings deep con­nec­tions to Don­ald Trump and con­tro­ver­sial Sil­i­con Val­ley ven­ture cap­i­tal­ist Peter Thiel.

Defense offi­cials announced Mon­day that Michael Krat­sios, the White House’s chief tech­nol­o­gy offi­cer, would serve as act­ing under­sec­re­tary for research and engi­neer­ing, a post that over­sees top-pri­or­i­ty projects in hyper­son­ics, quan­tum com­put­ing, micro­elec­tron­ics, and oth­er fields. He will con­tin­ue to serve in his White House role.

...

Krat­sios came to the White House in 2017 as deputy CTO, and moved up to CTO last year. He led efforts to fur­ther White House invest­ment in arti­fi­cial intel­li­gence and quan­tum sci­ence and to expand U.S. part­ner­ships in those areas. As the COVID-19 pan­dem­ic took hold, he helped launch a project to apply U.S. super­com­put­ers to the U.S response.

But Krat­sios was a “weird pick” for these senior tech­ni­cal roles, accord­ing to one per­son who has served as both a senior White House and Defense Depart­ment offi­cial advis­ing on tech­nol­o­gy issues.

Krat­sios grad­u­at­ed from Prince­ton with a bachelor’s degree in polit­i­cal sci­ence and a focus on ancient Greek democ­ra­cy. The per­son he’s replac­ing, Michael Grif­fin, holds a Ph.D. in aero­space engi­neer­ing and served as a NASA admin­is­tra­tor. Indeed, Krat­sios will be less aca­d­e­m­i­cal­ly cre­den­tialled than most of the pro­gram-man­agers he over­sees. So how did he get here?

After Prince­ton, he went to work for Peter Thiel, soon becom­ing CFO of Clar­i­um Cap­i­tal Man­age­ment, Thiel’s invest­ment com­pa­ny. He then became “chief of staff” for the tech bil­lion­aire, who was an ear­ly backer of the Trump cam­paign and who has played a key role in the administration’s approach to tech­nol­o­gy.

Thiel-backed ven­tures like Anduril and Palan­tir are play­ing a grow­ing role in the Defense Depart­ment. The for­mer offi­cial said the over­lap between Thiel-backed defense con­trac­tors and his pro­tege Krat­sios need not be a cause for con­cern. The Depart­ment has spent years try­ing to improve its rela­tion­ship with the pri­vate tech world from which Krat­sios emerged. But the offi­cial said Krat­sios might not prove to be the most effec­tive ambas­sador.

“It’s not clear to me that Krat­sios is warm­ing up Sil­i­con Val­ley,” the for­mer offi­cial said. “I don’t know how the rest of Sil­i­con Val­ley thinks of Krat­sios.”

Thiel has made a vari­ety of ene­mies in the tech world and beyond; for exam­ple, he has slammed Google as being too accom­mo­dat­ing to Chi­na.

The devel­op­ment, how­ev­er, is good news for “the Peter Thiel por­tion of Sil­i­con Val­ley,” the for­mer offi­cial said.

————

“Peter Thiel’s New Man In The Defense Depart­ment” by Patrick Tuck­er; Defense One; 07/13/2020

“Krat­sios grad­u­at­ed from Prince­ton with a bachelor’s degree in polit­i­cal sci­ence and a focus on ancient Greek democ­ra­cy. The per­son he’s replac­ing, Michael Grif­fin, holds a Ph.D. in aero­space engi­neer­ing and served as a NASA admin­is­tra­tor. Indeed, Krat­sios will be less aca­d­e­m­i­cal­ly cre­den­tialled than most of the pro­gram-man­agers he over­sees. So how did he get here?”

Yes, how exact­ly did Krat­sios get the job of the Pen­tagon’s top tech­nol­o­gy offi­cer despite hav­ing no dis­cernible tech­nol­o­gy exper­tise? He’s knows the right peo­ple. Specif­i­cal­ly Peter Thiel:

...
After Prince­ton, he went to work for Peter Thiel, soon becom­ing CFO of Clar­i­um Cap­i­tal Man­age­ment, Thiel’s invest­ment com­pa­ny. He then became “chief of staff” for the tech bil­lion­aire, who was an ear­ly backer of the Trump cam­paign and who has played a key role in the administration’s approach to tech­nol­o­gy.
...

But note how part of the sales pitch for Krat­sios get­ting this posi­tion is that he knows peo­ple in Sil­i­con Val­ley and that will help facil­i­tate rela­tion­ships between the Pen­ta­gon and Sil­i­con Val­ley firms. But as one for­mer offi­cial described it, it’s not like Krat­sios is actu­al­ly wide­ly like in Sil­i­con Val­ley in part due to his ties to Thiel and the fact that Thiel has cre­at­ed so many ene­mies. But Krat­sios’s selec­tion is unam­bigu­ous­ly good for “the Peter Thiel por­tion of Sil­i­con Val­ley.” And obvi­ous­ly obscene­ly good news for Thiel, who now has even more pow­er than ever. If you’re a Sil­i­con Val­ley firm that wants to do busi­ness with the Pen­ta­gon you had bet­ter not piss off Thiel:

...
Thiel-backed ven­tures like Anduril and Palan­tir are play­ing a grow­ing role in the Defense Depart­ment. The for­mer offi­cial said the over­lap between Thiel-backed defense con­trac­tors and his pro­tege Krat­sios need not be a cause for con­cern. The Depart­ment has spent years try­ing to improve its rela­tion­ship with the pri­vate tech world from which Krat­sios emerged. But the offi­cial said Krat­sios might not prove to be the most effec­tive ambas­sador.

“It’s not clear to me that Krat­sios is warm­ing up Sil­i­con Val­ley,” the for­mer offi­cial said. “I don’t know how the rest of Sil­i­con Val­ley thinks of Krat­sios.”

Thiel has made a vari­ety of ene­mies in the tech world and beyond; for exam­ple, he has slammed Google as being too accom­mo­dat­ing to Chi­na.

The devel­op­ment, how­ev­er, is good news for “the Peter Thiel por­tion of Sil­i­con Val­ley,” the for­mer offi­cial said.
...

Of course, the kind of pow­er wield­ed by Krat­sios is only going to last for as long as he’s the act­ing Under­sec­re­tary and that may now last long beyond the first months of 2021 if Trump isn’t reelect­ed. But any con­tracts set up could poten­tial­ly last much longer. In oth­er words, for Krat­sios and Thiel to ful­ly take advan­tage of this moment they are going to have to move fast and get as many long-term Pen­ta­gon con­tracts set up with Thiel-affil­i­at­ed firms as pos­si­ble.

So while the ascen­sion of Mike Grif­fin to the Under­sec­re­tary of Defense for Research and Engi­neer­ing served as a warn­ing that the defense acqui­si­tion process was going to be dra­mat­i­cal­ly sped up, it’s Griffin’s res­ig­na­tion that’s serv­ing as a warn­ing that this process could be kicked into over­drive.

And in prob­a­bly relat­ed news, guess which com­pa­ny just announced it’s going to be doing an IPO this year: yep, Palan­tir. It just announced it’s filed the IPO papers. So that’s going to be inter­est­ing to watch, espe­cial­ly with respect to how any new con­tracts that get announced this year might impact Palan­tir’s IPO val­u­a­tion. But as the fol­low­ing arti­cle describes, part of what this IPO announce­ment inter­est­ing is that it means Palan­tir is going to have to be more open to the pub­lic than before over the types of con­tracts it has with clients. Clients that include gov­ern­ments:

CNN

A secre­tive and con­tro­ver­sial start­up may go pub­lic. Here’s what you should know about it

By Rachel Metz and Sara Ash­ley O’Brien, CNN Busi­ness

Updat­ed 7:34 AM ET, Fri July 10, 2020

(CNN)In the 17 years since it was found­ed, Palan­tir Tech­nolo­gies has received finan­cial back­ing from the CIA, become one of the most valu­able pri­vate com­pa­nies in the US, and earned a seat at the table along­side the biggest tech com­pa­nies in meet­ing with Pres­i­dent Don­ald Trump.

But the Sil­i­con Val­ley data-ana­lyt­ics com­pa­ny, which is known for tak­ing on con­tro­ver­sial work for the US gov­ern­ment, has long shroud­ed itself in secre­cy — in part, as CEO and cofounder Alex Karp recent­ly explained in an inter­view with Axios on HBO, because clients often require them to keep qui­et. Now, it may have to shed at least some light on its oper­a­tions as the com­pa­ny inch­es clos­er to mak­ing a long-rumored Wall Street debut.

Palan­tir said this week that it con­fi­den­tial­ly filed paper­work with the US Secu­ri­ties and Exchange Com­mis­sion to go pub­lic. As with any pub­licly-trad­ed com­pa­ny, Palan­tir would need to dis­close more of its finan­cial his­to­ry and open itself to investor scruti­ny. And as with any tech com­pa­ny of its size — with a rough­ly $20 bil­lion val­u­a­tion — its ini­tial pub­lic offer­ing would like­ly be a high-pro­file event.

“If they do go down the IPO path, it would be a sem­i­nal moment in tech­nol­o­gy, very sim­i­lar to the likes of Uber and Lyft that were pri­vate for many years then went pub­lic,” said Daniel Ives, an ana­lyst with Wed­bush who tracks tech com­pa­nies.

Named after the see­ing stones in J.R.R. Tolkien’s fan­ta­sy nov­el “The Lord of the Rings,” Palan­tir is based in Palo Alto, Cal­i­for­nia. The com­pa­ny has two prod­ucts that cus­tomers use to orga­nize and glean insights from mounds of data: Gotham, which was ini­tial­ly devel­oped for gov­ern­ment clients, and Foundry. Any data con­tained in SEC fil­ings yet to be released pub­licly would almost cer­tain­ly reveal more.

With Palan­tir’s IPO, there’s more at stake than investor returns. Some hope an IPO would lead to greater trans­paren­cy, like Jac­in­ta Gon­za­lez, senior cam­paign orga­niz­er of orga­niz­ing group Mijente, which has long been crit­i­cal of Palan­tir for its work with US Immi­gra­tion and Cus­toms Enforce­ment, and its ties to the Trump admin­is­tra­tion. But she’s also wor­ried that a suc­cess­ful IPO will show more com­pa­nies that polic­ing is good for busi­ness.

Palan­tir spokes­woman Lisa Gor­don declined to com­ment for this sto­ry, cit­ing the qui­et peri­od Palan­tir entered as it began the IPO process.

A his­to­ry of bold claims—and con­tro­ver­sial work

Cofound­ed in 2003 by Peter Thiel, a mem­ber of the so-called “Pay­Pal mafia” and a long­time Face­book board mem­ber known for sup­port­ing Trump’s 2016 cam­paign, Palan­tir’s stat­ed mis­sion is to “make the West, espe­cial­ly Amer­i­ca, the strongest in the world.” The com­pa­ny touts its abil­i­ty to man­age and secure data at a mas­sive scale.

Palan­tir pro­vides gov­ern­ments and cor­po­ra­tions with tools to help with every­thing from track­ing the spread of the nov­el coro­n­avirus to zero­ing in on ter­ror­ists. It even report­ed­ly helped track down Osama bin Laden.

Yet it has his­tor­i­cal­ly been qui­et about the pre­cise ways in which its ser­vices are used, which has earned the com­pa­ny a shad­owy rep­u­ta­tion. What’s clear is that main­tain­ing a rela­tion­ship with the gov­ern­ment has been key to the com­pa­ny for much of its his­to­ry.

In its ear­ly years, Palan­tir strug­gled to get investors and clients, before rais­ing mon­ey from the CIA’s invest­ment arm, In-Q-Tel, as well as from Thiel and his VC firm Founders Fund. Since then, the com­pa­ny has tak­en on work with the US gov­ern­ment that oth­ers in Sil­i­con Val­ley might not have been com­fort­able with.

In 2017, CNN report­ed that Palan­tir had helped the Los Ange­les Police Depart­ment ana­lyze data, rang­ing from license plates pho­tos to rap sheets, traf­fic tick­ets, list­ings of fore­closed prop­er­ties and more. While this can make it eas­i­er for police to do things like track down crim­i­nals, it also indi­cates how tech­nol­o­gy such as Palan­tir’s offers law enforce­ment unprece­dent­ed sur­veil­lance capa­bil­i­ties.

More recent­ly, Karp said in an inter­view with CNBC at Davos in Jan­u­ary that Palan­tir has been assist­ing with find­ing undoc­u­ment­ed immi­grants for depor­ta­tion — some­thing that had long been rumored but uncon­firmed.

“We take what amounts to strong but often con­tro­ver­sial posi­tions,” Karp said at the time.

And in May, Karp acknowl­edged in an inter­view with Axios on HBO that Palan­tir’s tech­nol­o­gy “is used on occa­sion to kill peo­ple.” He made clear that it can be used for tar­get­ing of all kinds, includ­ing peo­ple. “If you’re look­ing for a ter­ror­ist in the world now you’re prob­a­bly using our gov­ern­ment prod­uct and you’re prob­a­bly doing the oper­a­tion that actu­al­ly takes out the per­son in anoth­er prod­uct we built,” he said, most like­ly refer­ring to Gotham and Foundry, respec­tive­ly.

At a time when com­pa­nies such as Ama­zon and Microsoft said they have halt­ed the sale of some con­test­ed tech­nol­o­gy — name­ly facial-recog­ni­tion sys­tems — to US police depart­ments, Palan­tir is “sort of the oppo­site,” said Evan Greer, deputy direc­tor at dig­i­tal rights non­prof­it Fight for the Future.

“I think com­pa­nies like Palan­tir are allow­ing gov­ern­ments and insti­tu­tions to weaponize our data and use it in ways that oppress peo­ple rather than lift them up,” Greer said.

The con­trar­i­an may cash in again

Just as Palan­tir is among the most con­tro­ver­sial com­pa­nies in Sil­i­con Val­ley, so is its most famous cofounder.

Thiel is known for being a con­trar­i­an. Thiel, a Stan­ford Uni­ver­si­ty grad­u­ate, funds a fel­low­ship that offers young entre­pre­neurs $100,000 to spend two years build­ing a com­pa­ny rather than attend­ing col­lege. He has argued that monop­o­lies are good and coau­thored a book in 1995 called “The Diver­si­ty Myth,” crit­i­ciz­ing the “debil­i­tat­ing impact” of “polit­i­cal­ly cor­rect mul­ti­cul­tur­al­ism” on col­lege edu­ca­tion.

He was also dis­cov­ered to be bankrolling the law­suit of Ter­ry Bol­lea, pop­u­lar­ly known as the wrestler Hulk Hogan, against Gawk­er Media that forced the pub­lish­er into bank­rupt­cy. Yet he has also donat­ed to the Com­mit­tee to Pro­tect Jour­nal­ists.

“It’s pre­cise­ly because I respect jour­nal­ists that I do not believe they are endan­gered by fight­ing back against Gawk­er,” he said in a 2016 inter­view, as part of his expla­na­tion for why he did­n’t think the two activ­i­ties con­tra­dict­ed each oth­er.

A lib­er­tar­i­an, he served as a sur­ro­gate for Don­ald Trump’s 2016 cam­paign, going against many pow­er­ful tech CEOs and crit­i­ciz­ing Sil­i­con Val­ley in the process.

Thiel, who has invest­ed at least $40 mil­lion in Palan­tir, has defend­ed the com­pa­ny’s role despite his pre­vi­ous­ly artic­u­lat­ed anti-gov­ern­ment approach. “The gov­ern­ment was col­lect­ing a lot of data [in the war on ter­ror­ism], more than they could ana­lyze,” he told For­tune in March 2016.. “If we could help them make sense of data, they could end indis­crim­i­nate sur­veil­lance.”

Unlike Pay­Pal, which was inspired by the idea of cre­at­ing a cur­ren­cy “free from all gov­ern­ment con­trol and dilu­tion,” Palan­tir is help­ing empow­er gov­ern­ment agen­cies: it scored $1.5 bil­lion in new con­tracts with the US gov­ern­ment in 2019 alone.

...

———–

“A secre­tive and con­tro­ver­sial start­up may go pub­lic. Here’s what you should know about it” by Rachel Metz and Sara Ash­ley O’Brien; CNN Busi­ness; 07/10/2020

“Palan­tir said this week that it con­fi­den­tial­ly filed paper­work with the US Secu­ri­ties and Exchange Com­mis­sion to go pub­lic. As with any pub­licly-trad­ed com­pa­ny, Palan­tir would need to dis­close more of its finan­cial his­to­ry and open itself to investor scruti­ny. And as with any tech com­pa­ny of its size — with a rough­ly $20 bil­lion val­u­a­tion — its ini­tial pub­lic offer­ing would like­ly be a high-pro­file event.”

It’s quite a con­ver­gence of events: Thiel gets Krat­sios installed in the per­fect posi­tion to shov­el all sorts of Pen­ta­gon con­tracts at Palan­tir right at the same time Palan­tir files an IPO. And there’s poten­tial­ly just a months left in the Trump admin­is­tra­tion so they have to move fast. And yet in order for this IPO to hap­pen Palan­tir needs to open itself up to investor scruti­ny to a degree it’s nev­er had to deal with before. What kind of hor­ri­ble secrets will be revealed? And will those hor­ri­ble secrets actu­al­ly harm Palan­tir’s per­ceived val­u­a­tion? It’s a defense con­trac­tor, after all. Hor­ri­ble secrets might be seen as an investor perk if they’re prof­itable hor­ri­ble secrets. So there’s plen­ty of ques­tions raised by the prospect of an Palan­tir IPO tak­ing place right when Thiel’s chief of staff because the new Pen­ta­gon head of tech­nol­o­gy pro­cure­ment, includ­ing the ques­tion of what hor­ri­ble com­pa­ny Peter Thiel is going to start next with all that new mon­ey he’s about to make.

Here’s a ‘good news’/‘bad news’ pair of sto­ries relat­ed to the Jan­u­ary 6 storm­ing of the Capi­tol and the sub­se­quent inves­ti­ga­tion into the iden­ti­ties of peo­ple in that insur­rec­tionary mob:

First, here’s a sto­ry from a cou­ple of weeks ago that points towards one of the good news aspects of this sto­ry. The sto­ry is about a wrong­ful arrest law­suit emerg­ing from a case where facial recog­ni­tion AI was used to iden­ti­fy a sus­pect. The man suing for wrong­ful arrest, Nijeer Parks, is Asian, and as the arti­cle notes, stud­ies have repeat­ed­ly shown that facial recog­ni­tion soft­ware does not per­form as well on Black and Asian faces. In Feb­ru­ary 2019, Nijeer Parks was accused of shoplift­ing can­dy and try­ing to hit a police offi­cer with a car at a Hamp­ton Inn in New Jer­sey.

Curi­ous­ly, there’s also a mys­tery as to who con­duct­ed the facial recog­ni­tion that wrong­ful­ly fin­gered Park. Park­s’s ini­tial law­suit accused Clearview AI of run­ning the face match search for the New Jer­sey Police. Recall how Clearview is the extreme­ly con­tro­ver­sial pri­vate facial recog­ni­tion com­pa­ny that appears to have scraped vir­tu­al­ly all of the pub­licly avail­able inter­net to amass a vast data­base of bil­lions of pho­tographs. Also recall how Clearview’s investors include Peter Thiel and the first appears to have close ties to the far right and the Repub­li­can Par­ty.

But there’s a ques­tion as to whether or not Clearview’s tools were actu­al­ly used in Park­s’s case. His attor­ney said he based the con­clu­sion that Clearview AI did the match based on pre­vi­ous reports that New Jer­sey law enforce­ment was already work­ing with Clearview to pro­vide these ser­vices. But Clearview denies that its soft­ware was used for the match. And accord­ing to the police report of Park­s’s arrest, the match was to a license pho­to, which would reside in a gov­ern­ment data­base that Clearview AI tech­ni­cal­ly can­not access. And yet the state agen­cies asked to run the face recog­ni­tion search — the New York State Intel­li­gence Cen­ter, New Jersey’s Region­al Oper­a­tions Intel­li­gence Cen­ter — said they did not make the match. What’s going on here? Is Clearview AI get­ting access to state data­bas­es of license pho­to infor­ma­tion its not sup­posed to have? We don’t know at this point. But it’s becom­ing increas­ing­ly clear that Clearview AI’s rela­tion­ship with US law enforce­ment is deep­en­ing.

So what’s the good news here? Well, the good news is that, should AI facial recog­ni­tion tech­nol­o­gy be used on the pro-Trump mob of peo­ple, at least it was a pre­dom­i­nant­ly white mob. And that means the exist­ing facial recog­ni­tion algo­rithms should prob­a­bly be a lot more accu­rate, mak­ing it much less like­ly that inno­cent peo­ple will get erro­neous­ly accused and face the same kind of night­mare Nijeer Parks faced based on bad facial recog­ni­tion soft­ware:

The New York Times

Anoth­er Arrest, and Jail Time, Due to a Bad Facial Recog­ni­tion Match

A New Jer­sey man was accused of shoplift­ing and try­ing to hit an offi­cer with a car. He is the third known Black man to be wrong­ful­ly arrest­ed based on face recog­ni­tion.

By Kash­mir Hill
Pub­lished Dec. 29, 2020
Updat­ed Jan. 6, 2021

In Feb­ru­ary 2019, Nijeer Parks was accused of shoplift­ing can­dy and try­ing to hit a police offi­cer with a car at a Hamp­ton Inn in Wood­bridge, N.J. The police had iden­ti­fied him using facial recog­ni­tion soft­ware, even though he was 30 miles away at the time of the inci­dent.

Mr. Parks spent 10 days in jail and paid around $5,000 to defend him­self. In Novem­ber 2019, the case was dis­missed for lack of evi­dence.

Mr. Parks, 33, is now suing the police, the pros­e­cu­tor and the City of Wood­bridge for false arrest, false impris­on­ment and vio­la­tion of his civ­il rights.

He is the third per­son known to be false­ly arrest­ed based on a bad facial recog­ni­tion match. In all three cas­es, the peo­ple mis­tak­en­ly iden­ti­fied by the tech­nol­o­gy have been Black men.

Facial recog­ni­tion tech­nol­o­gy is known to have flaws. In 2019, a nation­al study of over 100 facial recog­ni­tion algo­rithms found that they did not work as well on Black and Asian faces. Two oth­er Black men — Robert Williams and Michael Oliv­er, who both live in the Detroit area — were also arrest­ed for crimes they did not com­mit based on bad facial recog­ni­tion match­es. Like Mr. Parks, Mr. Oliv­er sued over the wrong­ful arrest.

Nathan Freed Wessler, an attor­ney with the Amer­i­can Civ­il Lib­er­ties Union who believes that the police should stop using face recog­ni­tion tech­nol­o­gy, said the three cas­es demon­strat­ed “how this tech­nol­o­gy dis­pro­por­tion­ate­ly harms the Black com­mu­ni­ty.”

“Mul­ti­ple peo­ple have now come for­ward about being wrong­ful­ly arrest­ed because of this flawed and pri­va­cy-invad­ing sur­veil­lance tech­nol­o­gy,” Mr. Wessler said. He wor­ries that there have been oth­er arrests and even mis­tak­en con­vic­tions that have not been uncov­ered.

Law enforce­ment often defends the use of facial recog­ni­tion, despite its flaws, by say­ing it is used only as a clue in a case and will not lead direct­ly to an arrest. But Mr. Parks’s expe­ri­ence is anoth­er exam­ple of an arrest based almost sole­ly on a sug­gest­ed match by the tech­nol­o­gy.

The Crime

On a Sat­ur­day in Jan­u­ary 2019, two police offi­cers showed up at the Hamp­ton Inn in Wood­bridge after receiv­ing a report about a man steal­ing snacks from the gift shop.

The alleged shoplifter — a Black man, near­ly 6 feet tall, wear­ing a black jack­et — was vis­it­ing a Hertz office in the hotel lob­by, try­ing to get the rental agree­ment for a gray Dodge Chal­lenger extend­ed. The offi­cers con­front­ed him, and he apol­o­gized, accord­ing to the police report. He said he would pay for the snacks and gave the offi­cers a Ten­nessee driver’s license.

When the offi­cers checked the license, they dis­cov­ered it was fraud­u­lent. Accord­ing to a police report, one of the offi­cers spot­ted a “big bag of sus­pect­ed mar­i­jua­na” in the man’s pock­et. They tried to hand­cuff him. That was when the man ran, los­ing a shoe on the way to his rental car, police said.

As he drove off, the man hit a parked police car and a col­umn in front of the hotel, the police said. One of the offi­cers said he had to jump out of the way to avoid being hit. The rental car was lat­er found aban­doned in a park­ing lot a mile away.

The Match

A detec­tive in the Wood­bridge Police Depart­ment sent the pho­to from the fake driver’s license to state agen­cies that had access to face recog­ni­tion tech­nol­o­gy, accord­ing to a police report.

The next day, state inves­ti­ga­tors said they had a facial recog­ni­tion match: Nijeer Parks, who lived in Pater­son, N.J., 30 miles away, and worked at a gro­cery store. The detec­tive com­pared Mr. Parks’s New Jer­sey state ID with the fake Ten­nessee driver’s license and agreed it was the same per­son. After a Hertz employ­ee con­firmed that the license pho­to was of the shoplifter, the police issued a war­rant for Mr. Parks’s arrest.

“I don’t think he looks like me,” Mr. Parks said. “The only thing we have in com­mon is the beard.”

Mr. Parks’s mis­tak­en arrest was first report­ed by NJ Advance Media, which said the facial recog­ni­tion app Clearview AI had been used in the case, based on a claim in Mr. Parks’s law­suit. His lawyer, Daniel Sex­ton, said he had inferred that Clearview AI was used, giv­en media reports about facial recog­ni­tion in New Jer­sey, but now believes he was mis­tak­en.

Clearview AI is a facial recog­ni­tion tool that uses bil­lions of pho­tos scraped from the pub­lic web, includ­ing Face­book, LinkedIn and Insta­gram. Clearview AI’s founder, Hoan Ton-That, said offi­cers affil­i­at­ed with the state agen­cies where infor­ma­tion was ana­lyzed in the case, known as fusion cen­ters, were not using his company’s app at that time.

Accord­ing to the police report, the match in this case was to a license pho­to, which would reside in a gov­ern­ment data­base, to which Clearview AI does not have access.

The state agen­cies asked to run the face recog­ni­tion search — the New York State Intel­li­gence Cen­ter, New Jersey’s Region­al Oper­a­tions Intel­li­gence Cen­ter — said they did not make the match. Two inves­ti­ga­tors who iden­ti­fied Mr. Parks with facial recog­ni­tion, at the Rock­land Coun­ty Sheriff’s Office and Pal­isades Inter­state Park­way Police, have not respond­ed to requests for com­ment. It is unclear how the Wood­bridge Police Department’s request was sent to those two inves­ti­ga­tors.

In Jan­u­ary, after a New York Times arti­cle about Clearview AI, New Jersey’s attor­ney gen­er­al, Gur­bir S. Gre­w­al, put a mora­to­ri­um on Clearview’s use by the police and announced an inves­ti­ga­tion into “this prod­uct or prod­ucts like it.” A spokesman for the attor­ney general’s office said that New Jersey’s Divi­sion of Crim­i­nal Jus­tice was still eval­u­at­ing the use of facial recog­ni­tion prod­ucts in the state, and that the devel­op­ment of a pol­i­cy gov­ern­ing their use was ongo­ing.

‘I Was Afraid’

After his arrest, Mr. Parks was held for 10 days at the Mid­dle­sex Coun­ty Cor­rec­tions Cen­ter. New Jersey’s no-bail sys­tem uses an algo­rithm that eval­u­ates the defendant’s risk rather than mon­ey to deter­mine whether a defen­dant can be released before tri­al.

A decade ago, Mr. Parks was arrest­ed twice and incar­cer­at­ed for sell­ing drugs. He was released in 2016. The pub­lic safe­ty assess­ment score he received, which would have tak­en his past con­vic­tions into account, was high enough that he was not released after his first hear­ing. His moth­er and fiancée hired an attor­ney, who was able to get him out of jail and into a pre­tri­al mon­i­tor­ing pro­gram.

His his­to­ry with the crim­i­nal jus­tice sys­tem is what made this inci­dent so scary, he said, because this would have been his third felony, mean­ing he was at risk of a long sen­tence. When the pros­e­cu­tor offered a plea deal, he almost took it even though he was inno­cent.

“I sat down with my fam­i­ly and dis­cussed it,” Mr. Parks said. “I was afraid to go to tri­al. I knew I would get 10 years if I lost.”

Mr. Parks was able to get proof from West­ern Union that he had been send­ing mon­ey at a phar­ma­cy in Hale­don, N.J., when the inci­dent hap­pened. At his last court hear­ing, he told the judge that he was will­ing to go to tri­al to defend him­self. But a few months lat­er, his case was dis­missed.

Robert Hub­n­er, the chief of the Wood­bridge Police Depart­ment, declined to com­ment on the case because of the pend­ing law­suit, but said his depart­ment had not been served the com­plaint. The Mid­dle­sex Coun­ty prosecutor’s office also declined to com­ment.

...

————–

“Anoth­er Arrest, and Jail Time, Due to a Bad Facial Recog­ni­tion Match” by Kash­mir Hill; The New York Times; 12/29/2020

“Facial recog­ni­tion tech­nol­o­gy is known to have flaws. In 2019, a nation­al study of over 100 facial recog­ni­tion algo­rithms found that they did not work as well on Black and Asian faces. Two oth­er Black men — Robert Williams and Michael Oliv­er, who both live in the Detroit area — were also arrest­ed for crimes they did not com­mit based on bad facial recog­ni­tion match­es. Like Mr. Parks, Mr. Oliv­er sued over the wrong­ful arrest.”

Over and over, stud­ies keep find­ing that facial recog­ni­tion soft­ware does­n’t work as well on non-white faces. And that means there’s inevitably going to be a lot more law­suits over facial recog­ni­tion-dri­ven wrong­ful arrests. In the case of Nijeer Parks, it appeared Clearview AI was the com­pa­ny that gen­er­ate the wrong match, but as the case has unfold­ed the ques­tion of who actu­al­ly cre­at­ed the match remains an open ques­tion. The wrong match appears to have come from no where:

...
Mr. Parks’s mis­tak­en arrest was first report­ed by NJ Advance Media, which said the facial recog­ni­tion app Clearview AI had been used in the case, based on a claim in Mr. Parks’s law­suit. His lawyer, Daniel Sex­ton, said he had inferred that Clearview AI was used, giv­en media reports about facial recog­ni­tion in New Jer­sey, but now believes he was mis­tak­en.

Clearview AI is a facial recog­ni­tion tool that uses bil­lions of pho­tos scraped from the pub­lic web, includ­ing Face­book, LinkedIn and Insta­gram. Clearview AI’s founder, Hoan Ton-That, said offi­cers affil­i­at­ed with the state agen­cies where infor­ma­tion was ana­lyzed in the case, known as fusion cen­ters, were not using his company’s app at that time.

Accord­ing to the police report, the match in this case was to a license pho­to, which would reside in a gov­ern­ment data­base, to which Clearview AI does not have access.

The state agen­cies asked to run the face recog­ni­tion search — the New York State Intel­li­gence Cen­ter, New Jersey’s Region­al Oper­a­tions Intel­li­gence Cen­ter — said they did not make the match. Two inves­ti­ga­tors who iden­ti­fied Mr. Parks with facial recog­ni­tion, at the Rock­land Coun­ty Sheriff’s Office and Pal­isades Inter­state Park­way Police, have not respond­ed to requests for com­ment. It is unclear how the Wood­bridge Police Department’s request was sent to those two inves­ti­ga­tors.

In Jan­u­ary, after a New York Times arti­cle about Clearview AI, New Jersey’s attor­ney gen­er­al, Gur­bir S. Gre­w­al, put a mora­to­ri­um on Clearview’s use by the police and announced an inves­ti­ga­tion into “this prod­uct or prod­ucts like it.” A spokesman for the attor­ney general’s office said that New Jersey’s Divi­sion of Crim­i­nal Jus­tice was still eval­u­at­ing the use of facial recog­ni­tion prod­ucts in the state, and that the devel­op­ment of a pol­i­cy gov­ern­ing their use was ongo­ing.
...

Did Clearview get improp­er access to the New Jer­sey license data­base? Giv­ing the com­pa­ny access would have obvi­ous util­i­ty to New Jer­sey’s law enforce­ment so it’s not incon­ceiv­able that such access was giv­en. But if so, it’s a sign of how deep Clearview AI’s rela­tion­ship is get­ting with gov­ern­ment agencies...something per­haps not unex­pect­ed giv­en Peter Thiel’s invest­ment in the com­pa­ny and the obscene­ly close rela­tion­ship between gov­ern­ment agen­cies and Thiel’s Palan­tir.

So will Clearview AI’s tools be used to help iden­ti­fy the indi­vid­u­als who par­tic­pat­ed in the raid on the Capi­tol. We don’t know but it’s cer­tain­ly seems like a pos­si­bil­i­ty. And if so, at least we should­n’t have to be as wor­ried about mis­match­es.

But even if we assume the issue of acci­den­tal mis­match­es will be large­ly addressed when the match­ing is done on an over­whelm­ing­ly white crowd, there anoth­er form or mis­match that we should prob­a­bly keep in mind: missed match­es that arise from the exceed­ing­ly close rela­tion­ship between Clearview AI and the far right. The kind of rela­tion­ship that should raise seri­ous ques­tions about whether or not Clearview AI can be trust­ed to not run cov­er for its fel­low far right allies. As the fol­low­ing Buz­zFeed piece from back in March describes, Clearview AI has dif­fer­ent “com­pa­ny type” cat­e­gories of users for its search data­base. The cat­e­gories are “Gov­ern­ment”, “Bank”, and “Investor”, etc. But there’s also a “Friend” cat­e­go­ry. And based on the doc­u­ments Buz­zFeed received, those “Friends” include com­pa­nies like SHW Part­ners LLC, a com­pa­ny found­ed by top Trump cam­paign offi­cial Jason Miller. And guess who turned out to be one of Clearview’s “test users”: Alt Right arch-troll Charles C. John­son. So while we haven’t yet seen any indi­ca­tion that Clearview AI is going to be used to iden­ti­fy by the insur­rec­tionary mob, and we haven’t seen any indi­ca­tion that Clearview AI is will to run cov­er for far right sus­pects, we’ve cer­tain­ly seen strong indi­ca­tions that Clearview is being used by law enforce­ment agen­cies and that the com­pa­ny has dis­turbing­ly close ties to the far right:

Buz­zFeed News

Secret Users Of Clearview AI’s Facial Recog­ni­tion Drag­net Includ­ed A For­mer Trump Staffer, A Troll, And Con­ser­v­a­tive Think Tanks

CEO Hoan Ton-That said his facial recog­ni­tion app is strict­ly for law enforce­ment. But he’s shared it with polit­i­cal con­nec­tions, poten­tial investors, and enti­ties des­ig­nat­ed as “Friend.”
Ryan Mac Buz­zFeed News Reporter
Car­o­line Hask­ins Buz­zFeed News Reporter
Logan McDon­ald Buz­zFeed Staff

Last updat­ed on March 25, 2020, at 1:16 p.m. ET
Post­ed on March 11, 2020, at 5:06 p.m. ET

On a flight to Boston in Jan­u­ary, James, a young aca­d­e­m­ic, caught the atten­tion of an aisle mate. “Do you go to school in the area?” the red-beard­ed man asked James, whose name has been changed to shield his iden­ti­ty, amid oth­er small talk about tech­nol­o­gy and sci­ence.

Lat­er, as the plane tax­ied to its gate at Boston Logan Inter­na­tion­al Air­port, the man sug­gest­ed the two con­nect on social media. But before James could agree or even give his full name, his fel­low pas­sen­ger pulled out his phone and asked to show him some­thing. He opened an app, snapped a pho­to of James, and then showed him the phone. The app, which flashed a ban­ner remind­ing its users not to talk to jour­nal­ists, pop­u­lat­ed with a hand­ful of pic­tures from James’ brother’s Insta­gram account. And James’ face was in every sin­gle one.

James was dumb­found­ed to see him­self in so many pho­tos on a stranger’s phone. Some of his social media accounts had been set to pri­vate, and he had been care­ful not to post too many pho­tos of him­self to the web. But the mys­te­ri­ous app had found him on his brother’s Insta­gram account with­in sec­onds.

Lat­er that evening, James received a friend request. It was from the beard­ed man on the plane — Charles C. John­son, a con­tro­ver­sial right-wing activist and accused Holo­caust denier with ties to the Trump admin­is­tra­tion. John­son did not respond to mul­ti­ple requests for com­ment. (Fol­low­ing the pub­li­ca­tion of this sto­ry, John­son denied accu­sa­tions of Holo­caust denial.)

After read­ing cov­er­age about a new facial recog­ni­tion tool, James deduced that John­son had iden­ti­fied him using Clearview AI, a secre­tive com­pa­ny that’s claimed to have scraped more than 3 bil­lion pho­tos from social media and the web. Last month, a Buz­zFeed News inves­ti­ga­tion found that peo­ple at more than 2,200 orga­ni­za­tions have tried Clearview’s facial recog­ni­tion tech­nol­o­gy, includ­ing fed­er­al enti­ties such as Immi­gra­tion and Cus­toms Enforce­ment, the FBI, and pri­vate com­pa­nies like Macy’s, the NBA, and Bank of Amer­i­ca.

Pri­or to Buz­zFeed News’ report, Clearview insist­ed its tool was strict­ly for law enforce­ment. “Clearview AI’s search engine is avail­able only for law enforce­ment agen­cies and select secu­ri­ty pro­fes­sion­als to use as an inves­tiga­tive tool, and its results con­tain only pub­lic infor­ma­tion,” the com­pa­ny wrote in a Jan. 27 blog post.. “Accord­ing­ly, the Clearview app has built-in safe­guards to ensure these trained pro­fes­sion­als only use it for its intend­ed pur­pose: to help iden­ti­fy the per­pe­tra­tors and vic­tims of crimes.”

Clearview, how­ev­er, has shared its tech­nol­o­gy with orga­ni­za­tions it des­ig­nat­ed as friends, con­ser­v­a­tive think tanks, Repub­li­can law­mak­ers, and more than 20 poten­tial investors around the world, accord­ing to com­pa­ny doc­u­ments seen by Buz­zFeed News. Some of those enti­ties have con­nec­tions to the far right and the Trump admin­is­tra­tion, as do Clearview CEO Hoan Ton-That, a MAGA-sup­port­ing mobile app devel­op­er, and cofounder Richard Schwartz, who was once an advis­er to for­mer New York City may­or Rudy Giu­liani.

...

The wide­spread, unreg­u­lat­ed use of Clearview’s tech­nol­o­gy has con­cerned both Demo­c­ra­t­ic and Repub­li­can law­mak­ers. Last week, the House Com­mit­tee on Sci­ence, Space, and Tech­nol­o­gy sent a let­ter signed by each party’s rank­ing mem­ber to Ton-That ask­ing how his com­pa­ny deter­mined who gets access to its prod­ucts. That was fol­lowed by a sep­a­rate note to Clearview from Sen. Ed Markey of Mass­a­chu­setts, who sought to under­stand why the start­up had pro­vid­ed its soft­ware to orga­ni­za­tions out­side of law enforce­ment and to enti­ties in coun­tries with his­to­ries of human rights vio­la­tions, includ­ing Sau­di Ara­bia and the Unit­ed Arab Emi­rates..

Clearview — which is fac­ing mul­ti­ple law­suits from con­cerned cit­i­zens and poten­tial legal action from tech giants includ­ing Face­book and Google for scrap­ing their images — has main­tained in media inter­views that “it’s strict­ly for law enforce­ment.” But com­pa­ny data viewed by Buz­zFeed News appeared to con­tra­dict that state­ment.

A Buz­zFeed News review of the pub­licly avail­able code for Clearview’s web app found that the com­pa­ny had cre­at­ed a list of “com­pa­ny type” des­ig­na­tions — includ­ing “Gov­ern­ment,” “Bank,” and “Investor” — for enti­ties that had access to its tool. Among those 12 labels was one for “Friend.” A secu­ri­ty researcher on Twit­ter also dis­cov­ered the same label in an inde­pen­dent review of the code.

On a list of more than 2,200 enti­ties seen by Buz­zFeed News, Clearview appar­ent­ly des­ig­nat­ed the “com­pa­ny type” of a hand­ful of orga­ni­za­tions as “Friend.” Among them was SHW Part­ners LLC, a com­pa­ny found­ed by Jason Miller, a for­mer Trump cam­paign senior com­mu­ni­ca­tions offi­cial and one-time nom­i­nee for White House com­mu­ni­ca­tions direc­tor.

Miller, who cohosts a pod­cast with Trump’s for­mer chief strate­gist Steve Ban­non, declined to com­ment on Clearview or why his firm was list­ed as hav­ing access to the company’s facial recog­ni­tion soft­ware. Clearview’s records show that an account tied to SHW had run near­ly 20 search­es, some as recent­ly as Octo­ber.

Buz­zFeed News pre­vi­ous­ly revealed that Clearview’s data list­ed the offices of four Repub­li­can mem­bers of Con­gress — includ­ing Rep. John Rat­cliffe, a cur­rent nom­i­nee for the direc­tor of nation­al intel­li­gence — as hav­ing been giv­en accounts, along with some­one asso­ci­at­ed with the “White House Tech Office.” That White House–affiliated account was cre­den­tialed in Sep­tem­ber 2019 and per­formed six search­es.

“If a cur­rent or for­mer staff mem­ber attempt­ed to access more infor­ma­tion about this prod­uct, it was not an offi­cial inquiry and was not sanc­tioned by the White House,” a senior White House offi­cial told Buz­zFeed News last month.

Beyond SHW, oth­er enti­ties on Clearview’s list with the “Friend” des­ig­na­tion includ­ed the Samar­i­an Group, a New York–based pri­vate equi­ty firm; Droese Raney, a Dal­las com­mer­cial archi­tec­ture com­pa­ny; and Tor Eke­land, Clearview’s out­side law firm, whose name was mis­spelled “Tor Eck­lund” on the list. All three orga­ni­za­tions did not respond to requests for com­ment.

Clearview also pro­vid­ed access to two con­ser­v­a­tive think tanks: the Man­hat­tan Insti­tute and the Amer­i­can Enter­prise Insti­tute.

The Man­hat­tan Insti­tute, where cofounders Ton-That and Schwartz report­ed­ly first met, received an account in the fall. A spokesper­son for the think tank said Schwartz is a “long­time friend of the Insti­tute” and not­ed that a for­mer fel­low at the orga­ni­za­tion had served as one of three judges for a sup­pos­ed­ly inde­pen­dent report in which Clearview claimed its tech­nol­o­gy was “100% accu­rate.”

Although the spokesper­son denied that the Man­hat­tan Insti­tute used Clearview, a doc­u­ment reviewed by Buz­zFeed News shows that one per­son asso­ci­at­ed with the orga­ni­za­tion ran more than a dozen search­es.

An indi­vid­ual asso­ci­at­ed with the Amer­i­can Enter­prise Insti­tute is also list­ed with an account, which, accord­ing to Clearview doc­u­ments seen by Buz­zFeed News, was designed to dis­play a ban­ner read­ing “Richard Says Hi” — an appar­ent ref­er­ence to Schwartz. AEI did not return mul­ti­ple requests for com­ment.

While the data seen by Buz­zFeed News list­ed thou­sands of orga­ni­za­tions with accounts, there were oth­er unknown indi­vid­ual users who ran search­es with Clearview’s app. On its list, the start­up main­tained an entry for “Clearview Test Users,” which includ­ed more than 220 accounts that had run more than 30,000 search­es over a 19-month peri­od.

When asked if John­son was one of its test users, Clearview, whose founders have a his­to­ry of con­nec­tions to indi­vid­u­als on the far right, declined to answer. John­son has known Ton-That since at least 2016, when the two were pho­tographed at a din­ner togeth­er.

...

Based on a list of cre­den­tialed users reviewed by Buz­zFeed News, Clearview met with indi­vid­u­als from many of Sil­i­con Valley’s most notable firms, among them Klein­er Perkins and Grey­lock Part­ners. A Grey­lock Part­ners spokesper­son said a firm staffer met Ton-That at “a defense indus­try din­ner” in late 2018 and was giv­en a demo account that was sparse­ly used. Grey­lock is not an investor in the com­pa­ny, they said. A Klein­er Perkins spokesper­son did not respond to an email request for com­ment.

Oth­er firms did not invest but are list­ed in Clearview doc­u­ments as hav­ing run hun­dreds of search­es. Among them was ven­ture cap­i­tal firm Data Col­lec­tive Ven­ture Cap­i­tal, which ran more than 270 scans, some as recent­ly as last month. A spokesper­son for the firm con­firmed it had used Clearview, but not­ed that “the prod­uct nev­er made it past our pre­lim­i­nary due dili­gence phase.”

“DCVC invests in Deep Tech com­pa­nies includ­ing those in the AI pri­va­cy and secu­ri­ty space,” they said. “Through pre­lim­i­nary mar­ket due dili­gence, we eval­u­at­ed the appli­ca­tion with test search­es of friends and fam­i­ly, con­duct­ed with their full con­sent.”

Pass­port Cap­i­tal, a San Francisco–based firm which did not respond to a request for com­ment, ran more than 350 search­es, accord­ing to doc­u­ments seen by Buz­zFeed News. Anoth­er account linked to Sequoia Cap­i­tal was list­ed as hav­ing run more than 210 scans over a 10-month peri­od. The New York Times pre­vi­ous­ly report­ed that Sequoia man­ag­ing part­ner Doug Leone had used Clearview.

“We don’t com­ment on com­pa­nies we pass on,” a Sequoia spokesper­son told Buz­zFeed News. “We reg­u­lar­ly test apps while we are eval­u­at­ing an invest­ment, and we no longer have access to this app.”

Founders Fund, the ven­ture firm found­ed by Peter Thiel, a bil­lion­aire and Face­book board mem­ber, also appears on Clearview’s list. There were two accounts asso­ci­at­ed with the San Francisco–based ven­ture firm that have run about 70 search­es, some as recent­ly as Octo­ber, accord­ing to Clearview’s records. A com­pa­ny spokesper­son did not reply to requests for com­ment.

A spokesper­son for Thiel, who per­son­al­ly invest­ed $200,000 in Smartcheckr, Ton-That’s pre­de­ces­sor com­pa­ny — an invest­ment that was lat­er con­vert­ed into a stake in Clearview — did not pro­vide com­ment. Thiel’s per­son­al invest­ment firm, Thiel Cap­i­tal, had a Clearview login, though it appears it was nev­er used, based on doc­u­ments seen by Buz­zFeed News.

Clearview didn’t just stick to tra­di­tion­al ven­ture firms. Accord­ing to doc­u­ments seen by Buz­zFeed News, it pro­vid­ed two accounts to indi­vid­u­als with ties to Iconiq Cap­i­tal, an invest­ment firm that man­ages the mon­ey of Face­book CEO Mark Zucker­berg and oth­er tech bil­lion­aires. Those accounts ran more than 70 search­es, some as recent­ly as Feb­ru­ary. Iconiq did not respond to requests for com­ment.

Ocu­lus Rift cofounder Palmer Luck­ey, who now leads defense con­trac­tor Anduril Indus­tries, is also list­ed in Clearview’s data as hav­ing an account. A spokesper­son told Buz­zFeed News that Anduril “has nev­er had any rela­tion­ship of any kind” with the com­pa­ny. Clearview data shows that account as hav­ing run more than 20 search­es.

“Clearview AI, as part of its fundrais­ing efforts, con­tact­ed Palmer Luck­ey in his indi­vid­ual capac­i­ty via his per­son­al email to ask him if he would con­sid­er invest­ing in their com­pa­ny,” the spokesper­son said. “He was sent a test account with­out request­ing one.”

Beyond Sil­i­con Val­ley, Clearview also cre­den­tialed peo­ple asso­ci­at­ed with Japan­ese con­glom­er­ate Soft­Bank, and RIT Cap­i­tal Part­ners, a the UK-based investor in the start­up. Mubadala Invest­ment Com­pa­ny, a sov­er­eign wealth fund owned by the gov­ern­ment of Abu Dhabi in the Unit­ed Arab Emi­rates, is list­ed in the data as hav­ing six accounts that ran more than 100 search­es, some as recent­ly as Jan­u­ary. An RIT spokesper­son con­firmed it had made “a very small minor­i­ty invest­ment” in Clearview. Soft­Bank declined to answer ques­tions, and Mubadala did not respond to mul­ti­ple requests for com­ment.

Clearview data reviewed by Buz­zFeed News list­ed more than 20 ven­ture and pri­vate equi­ty firms the com­pa­ny had cre­den­tialed, but there were oth­ers as well. An entry titled “Clearview Investors” showed five asso­ci­at­ed accounts, though it’s unclear whom those accounts belonged to.

Among the known Clearview investors who did not appear in that data was Kire­na­ga Part­ners, an invest­ment firm based in Bronxville, New York, led by David Scal­zo. A vocal pro­po­nent of the start­up, Scal­zo has appeared on CNBC and var­i­ous pod­casts to defend the com­pa­ny and its tech­nol­o­gy.

One of those appear­ances was in a Jan­u­ary inter­view with Dil­bert cre­ator Scott Adams, who asked the investor if he thought “it’s inevitable the pub­lic will have [Clearview AI].”

“It is inevitable that this dig­i­tal infor­ma­tion will be out there,” Scal­zo replied.

————-

“Secret Users Of Clearview AI’s Facial Recog­ni­tion Drag­net Includ­ed A For­mer Trump Staffer, A Troll, And Con­ser­v­a­tive Think Tanks” by Ryan Mac, Car­o­line Hask­ins, and Logan McDon­ald; Buz­zFeed News; 03/11/2020

After read­ing cov­er­age about a new facial recog­ni­tion tool, James deduced that John­son had iden­ti­fied him using Clearview AI, a secre­tive com­pa­ny that’s claimed to have scraped more than 3 bil­lion pho­tos from social media and the web. Last month, a Buz­zFeed News inves­ti­ga­tion found that peo­ple at more than 2,200 orga­ni­za­tions have tried Clearview’s facial recog­ni­tion tech­nol­o­gy, includ­ing fed­er­al enti­ties such as Immi­gra­tion and Cus­toms Enforce­ment, the FBI, and pri­vate com­pa­nies like Macy’s, the NBA, and Bank of Amer­i­ca.

Clearview AI pre­vi­ous­ly claimed its tools were exclu­sive­ly for law enforce­ment. But Buz­zFeed found more the 2,200 enti­ties had used the tool. Includ­ing a dis­turb­ing num­ber of enti­ties and fig­ures asso­ci­at­ed with the Repub­li­can Par­ty and Trump White House. Jason Miller’s com­pa­ny was even giv­en “Friend” sta­tus:

...
Pri­or to Buz­zFeed News’ report, Clearview insist­ed its tool was strict­ly for law enforce­ment. “Clearview AI’s search engine is avail­able only for law enforce­ment agen­cies and select secu­ri­ty pro­fes­sion­als to use as an inves­tiga­tive tool, and its results con­tain only pub­lic infor­ma­tion,” the com­pa­ny wrote in a Jan. 27 blog post.. “Accord­ing­ly, the Clearview app has built-in safe­guards to ensure these trained pro­fes­sion­als only use it for its intend­ed pur­pose: to help iden­ti­fy the per­pe­tra­tors and vic­tims of crimes.”

Clearview, how­ev­er, has shared its tech­nol­o­gy with orga­ni­za­tions it des­ig­nat­ed as friends, con­ser­v­a­tive think tanks, Repub­li­can law­mak­ers, and more than 20 poten­tial investors around the world, accord­ing to com­pa­ny doc­u­ments seen by Buz­zFeed News. Some of those enti­ties have con­nec­tions to the far right and the Trump admin­is­tra­tion, as do Clearview CEO Hoan Ton-That, a MAGA-sup­port­ing mobile app devel­op­er, and cofounder Richard Schwartz, who was once an advis­er to for­mer New York City may­or Rudy Giu­liani.

...
A Buz­zFeed News review of the pub­licly avail­able code for Clearview’s web app found that the com­pa­ny had cre­at­ed a list of “com­pa­ny type” des­ig­na­tions — includ­ing “Gov­ern­ment,” “Bank,” and “Investor” — for enti­ties that had access to its tool. Among those 12 labels was one for “Friend.” A secu­ri­ty researcher on Twit­ter also dis­cov­ered the same label in an inde­pen­dent review of the code.

On a list of more than 2,200 enti­ties seen by Buz­zFeed News, Clearview appar­ent­ly des­ig­nat­ed the “com­pa­ny type” of a hand­ful of orga­ni­za­tions as “Friend.” Among them was SHW Part­ners LLC, a com­pa­ny found­ed by Jason Miller, a for­mer Trump cam­paign senior com­mu­ni­ca­tions offi­cial and one-time nom­i­nee for White House com­mu­ni­ca­tions direc­tor.

Miller, who cohosts a pod­cast with Trump’s for­mer chief strate­gist Steve Ban­non, declined to com­ment on Clearview or why his firm was list­ed as hav­ing access to the company’s facial recog­ni­tion soft­ware. Clearview’s records show that an account tied to SHW had run near­ly 20 search­es, some as recent­ly as Octo­ber.

Buz­zFeed News pre­vi­ous­ly revealed that Clearview’s data list­ed the offices of four Repub­li­can mem­bers of Con­gress — includ­ing Rep. John Rat­cliffe, a cur­rent nom­i­nee for the direc­tor of nation­al intel­li­gence — as hav­ing been giv­en accounts, along with some­one asso­ci­at­ed with the “White House Tech Office.” That White House–affiliated account was cre­den­tialed in Sep­tem­ber 2019 and per­formed six search­es.

“If a cur­rent or for­mer staff mem­ber attempt­ed to access more infor­ma­tion about this prod­uct, it was not an offi­cial inquiry and was not sanc­tioned by the White House,” a senior White House offi­cial told Buz­zFeed News last month.

...

Clearview also pro­vid­ed access to two con­ser­v­a­tive think tanks: the Man­hat­tan Insti­tute and the Amer­i­can Enter­prise Insti­tute.

...

Founders Fund, the ven­ture firm found­ed by Peter Thiel, a bil­lion­aire and Face­book board mem­ber, also appears on Clearview’s list. There were two accounts asso­ci­at­ed with the San Francisco–based ven­ture firm that have run about 70 search­es, some as recent­ly as Octo­ber, accord­ing to Clearview’s records. A com­pa­ny spokesper­son did not reply to requests for com­ment.

A spokesper­son for Thiel, who per­son­al­ly invest­ed $200,000 in Smartcheckr, Ton-That’s pre­de­ces­sor com­pa­ny — an invest­ment that was lat­er con­vert­ed into a stake in Clearview — did not pro­vide com­ment. Thiel’s per­son­al invest­ment firm, Thiel Cap­i­tal, had a Clearview login, though it appears it was nev­er used, based on doc­u­ments seen by Buz­zFeed News.
...

And then there’s the fact that Charles C. John­son appears to be a test user with full access to just run search­es when­ev­er he wants:

...
While the data seen by Buz­zFeed News list­ed thou­sands of orga­ni­za­tions with accounts, there were oth­er unknown indi­vid­ual users who ran search­es with Clearview’s app. On its list, the start­up main­tained an entry for “Clearview Test Users,” which includ­ed more than 220 accounts that had run more than 30,000 search­es over a 19-month peri­od.

When asked if John­son was one of its test users, Clearview, whose founders have a his­to­ry of con­nec­tions to indi­vid­u­als on the far right, declined to answer. John­son has known Ton-That since at least 2016, when the two were pho­tographed at a din­ner togeth­er.
...

Keep in mind that some­one like Charles C. John­son prob­a­bly per­son­al­ly knows a num­ber of the peo­ple who stormed the Capi­tol. That’s why his ties to Clearview are so poten­tial­ly so sig­nif­i­cant in this case.

So the good news is that con­tem­po­rary facial recog­ni­tion soft­ware should­n’t suf­fer from too much racial­ly biased inac­cu­rate match­es if applied to Trump’s Capi­tol mili­tia. The bad news is that the riot­ers are lit­er­al­ly going to be ‘friends of friends’ of the com­pa­ny that’s prob­a­bly doing the match­ing.

Worse than Water­gate? It’s one of the meta ques­tions for the Trump era that is once again being asked fol­low­ing the grow­ing rev­e­la­tions about the Trump Depart­ment of Jus­tice spy­ing on not just Demo­c­ra­t­ic mem­bers of con­gress but also their fam­i­ly mem­bers in a quest to find gov­ern­ment leak­ers. It’s the kind of sto­ry that rais­es ques­tions about who was­n’t being spied on by the Trump admines­tra­tion. So with ques­tions about secret gov­ern­ment spy­ing once again being asked, it’s worth keep­ing in mind one of the con­tem­po­rary con­texts of secret gov­ern­ment spy­ing oper­a­tions. In par­tic­u­lar spy­ing by Repub­li­can admin­is­tra­tions: much of the US’s nation­al secu­ri­ty ana­lyt­i­cal capa­bil­i­ties are being car­ried out by pri­vate enti­ties like Palan­tir. And since Palan­tir’s ser­vices to clients includes the iden­ti­fi­ca­tion of leak­ers, we can’t rule out the pos­si­bil­i­ty that the Trump admin­is­tra­tion was­n’t just task­ing the Depart­ment of Jus­tice in its leak hunt. A pri­vate enti­ty like Palan­tir would almost be ide­al for a scan­dalous oper­a­tion of that nature, espe­cial­ly for the Trump admin­is­tra­tion that ben­e­fit­ed from an extreme­ly close polit­i­cal alliance between Trump and Palan­tir co-founder Peter Thiel.

So was Palan­tir at all involved in this lat­est ‘worse than Watergate’-level Trump scan­dal? We have no idea. More impor­tant­ly, we have no idea if the ques­tion is even being asked by inves­ti­ga­tors. But as the fol­low­ing 2019 piece in Vice makes clear, Palan­tir was def­i­nite­ly inter­est­ed in offer­ing leak-hunt­ing ser­vices, the kind of ser­vice that was almost ide­al for work­ing with the Palan­tir Big Data mod­el of know­ing as much as pos­si­ble about as many peo­ple as pos­si­ble:

Vice
Moth­er­board

These Videos Show How Palan­tir Tracks Leak­ers, Pro­test­ers, and Pris­on­ers
Palantir’s pro­gram isn’t just used by law enforce­ment, but by third-par­ties who offer aug­ment­ed ver­sions.

by Edward Ong­we­so Jr
August 21, 2019, 9:25am

Palan­tir is a big data ana­lyt­ics com­pa­ny with a rep­u­ta­tion for being incred­i­bly secre­tive. That hasn’t helped it escape pub­lic scruti­ny for its col­lab­o­ra­tion with law enforce­ment agen­cies such as the North­ern Cal­i­for­nia Region­al Intel­li­gence Cen­ter, its crit­i­cal role as a tech­nol­o­gy back­bone for Immi­gra­tion and Cus­toms Enforce­ment (ICE), or the cozy rela­tion­ship between Don­ald Trump and its founder, Peter Thiel.

Palan­tir has a long his­to­ry of work­ing with third par­ties, includ­ing intel­li­gence agen­cies and select ven­dors known as “Pre­ferred Part­ners” that are autho­rized to pro­vide “select forms of prod­uct sup­port ser­vices.” On YouTube, there is a whole ecosys­tem of tuto­r­i­al videos by third-par­ty data ana­lyt­ics com­pa­nies adver­tis­ing all the things you can do with Palan­tir and sim­i­lar soft­ware, such as IBM’s i2 Enter­prise Insight Analy­sis.

One of those ven­dors is Prae­scient Ana­lyt­ics, an Alexan­dria, VA-based com­pa­ny. Prae­scient Ana­lyt­ics isn’t cur­rent­ly a Pre­ferred Part­ner but has pro­vid­ed Palan­tir sup­port ser­vices to law enforce­ment agen­cies such as the Cook Coun­ty Sheriff’s Depart­ment and the Los Ange­les Sheriff’s Depart­ment.

Through its pres­ence on YouTube, Prae­scient explains its com­mit­ment to “apply­ing cut­ting edge ana­lyt­ic tech­nolo­gies and method­olo­gies to sup­port gov­ern­ment and com­mer­cial clients.” For exam­ple, in one video, the com­pa­ny demon­strates how an orga­ni­za­tion can use Palan­tir’s soft­ware to find out if one of its employ­ees leaked con­fi­den­tial infor­ma­tion to a blog­ger.

The demon­stra­tion revolves around a car com­pa­ny that had to recall its prod­uct due to faulty wiring. The com­pa­ny con­duct­ed an inter­nal report on how this hap­pened and who was respon­si­ble for the defect, but the inter­nal report was leaked to the press via a 10-minute voice­mail. Prae­scient claims that this demon­stra­tion is based on a real case that it han­dled.

The idea of plug­ging an inter­nal leak is not new, but Praescient’s demon­stra­tion video shows how to eas­i­ly plug it with Palantir’s data. First, a user needs to search for emails sent out­side the com­pa­ny, then for phone calls made with­in a cer­tain peri­od that last­ed longer than 10 min­utes. This pro­duces a list of sus­pects and whether they made a call to an inter­nal or exter­nal num­ber. It also depicts each suspect’s posi­tion in the cor­po­rate struc­ture to deter­mine if they could’ve had access to the con­fi­den­tial infor­ma­tion. Prae­scien­t’s tools can then auto­mat­i­cal­ly pull social media data to find out if the sus­pects have pri­ma­ry or sec­ondary con­nec­tions to the blog­ger. If the sus­pect has a pri­vate pro­file, then Prae­scient can tar­get oth­er con­nec­tions in their social net­work and scrap those pro­files. After a con­nec­tion is dis­cov­ered, a clos­er inves­ti­ga­tion of the mutu­al con­nec­tion’s social media uncov­ers a pho­to where the sus­pect and the blog­ger are both present.

These sorts of meth­ods can be applied to a host of oth­er sce­nar­ios. In anoth­er video, the com­pa­ny claims the same tools can be used to track the polit­i­cal and mil­i­tary dimen­sions of Kim Jong Un’s ascen­sion to pow­er in North Korea.

In anoth­er video Prae­scient claims Palan­tir’s tech can help a user map out the con­nec­tions between var­i­ous peo­ple and neigh­bor­hoods involved in the 2011 UK riots. Most com­men­tary and analy­sis high­lights police bru­tal­i­ty as a major fac­tor in spark­ing the riots. The demon­stra­tor uses Palantir’s soft­ware to try and chal­lenge what the valid­i­ty of that nar­ra­tive, sug­gest­ing riot­ers were sim­ply “show[ing] the police we can do what we want” or oppor­tunis­tic loot­ers with con­nec­tions to orga­nized crime.

Anoth­er video shows how Prae­scient can be used for “Prison Man­age­ment Sup­port.” The soft­ware allows a user to track inmates in real-time, claim­ing that by allow­ing jailors to “store diverse sets of infor­ma­tion,” it can allow guards to “proac­tive­ly mon­i­tor inmates pro­files and under­stand their capa­bil­i­ties.” In oth­er words, pris­on­ers can be more effec­tive­ly seg­re­gat­ed or iso­lat­ed if the data ana­lyt­ics jus­ti­fies sus­pi­cion about where they work, who they asso­ciate with, or which cell block they are caged in.

Now that Palan­tir has renewed its con­tract with ICE through 2022 for over $49 mil­lion, it is impor­tant to try and shed light on how Palantir’s soft­ware is used—not just by ICE, but by third-par­ty ven­dors who pro­vide sup­port to the pri­vate sec­tor and law enforce­ment.

...

———–

“These Videos Show How Palan­tir Tracks Leak­ers, Pro­test­ers, and Pris­on­ers” by Edward Ong­we­so Jr; Vice Moth­er­board; 08/21/2019

“Through its pres­ence on YouTube, Prae­scient explains its com­mit­ment to “apply­ing cut­ting edge ana­lyt­ic tech­nolo­gies and method­olo­gies to sup­port gov­ern­ment and com­mer­cial clients.” For exam­ple, in one video, the com­pa­ny demon­strates how an orga­ni­za­tion can use Palan­tir’s soft­ware to find out if one of its employ­ees leaked con­fi­den­tial infor­ma­tion to a blog­ger.”

While we don’t have any direct evi­dence the Trump admin­is­tra­tion uti­lized Palan­tir’s leak-hunt­ing ser­vices, it seems high­ly like­ly the Trump admin­is­tra­tion was at least aware such ser­vices exist­ed. Which rais­es the ques­tion about whether or not the US gov­ern­ment was already uti­liz­ing these leak-hunt­ing ser­vices before this scan­dal even start­ed. The US gov­ern­ment is a major Palan­tir client and helped start the com­pa­ny in the first place, after all. In that con­text, it would almost be sur­pris­ing if these ser­vices weren’t be uti­lized by the US agen­cies:

...
The idea of plug­ging an inter­nal leak is not new, but Praescient’s demon­stra­tion video shows how to eas­i­ly plug it with Palantir’s data. First, a user needs to search for emails sent out­side the com­pa­ny, then for phone calls made with­in a cer­tain peri­od that last­ed longer than 10 min­utes. This pro­duces a list of sus­pects and whether they made a call to an inter­nal or exter­nal num­ber. It also depicts each suspect’s posi­tion in the cor­po­rate struc­ture to deter­mine if they could’ve had access to the con­fi­den­tial infor­ma­tion. Prae­scien­t’s tools can then auto­mat­i­cal­ly pull social media data to find out if the sus­pects have pri­ma­ry or sec­ondary con­nec­tions to the blog­ger. If the sus­pect has a pri­vate pro­file, then Prae­scient can tar­get oth­er con­nec­tions in their social net­work and scrap those pro­files. After a con­nec­tion is dis­cov­ered, a clos­er inves­ti­ga­tion of the mutu­al con­nec­tion’s social media uncov­ers a pho­to where the sus­pect and the blog­ger are both present.

These sorts of meth­ods can be applied to a host of oth­er sce­nar­ios. In anoth­er video, the com­pa­ny claims the same tools can be used to track the polit­i­cal and mil­i­tary dimen­sions of Kim Jong Un’s ascen­sion to pow­er in North Korea.

...

Now that Palan­tir has renewed its con­tract with ICE through 2022 for over $49 mil­lion, it is impor­tant to try and shed light on how Palantir’s soft­ware is used—not just by ICE, but by third-par­ty ven­dors who pro­vide sup­port to the pri­vate sec­tor and law enforce­ment.
...

And that’s why one of the big ques­tions sur­round­ing this sto­ry is whether or not ques­tions about Palan­tir’s poten­tial involve­ment are being asked at all. Palan­tir is an obvi­ous sus­pect for any Trump-relat­ed Big Data abuse scan­dal. Per­haps the obvi­ous sus­pect. And yet the US gov­ern­men­t’s rela­tion­ship with Palan­tir is also obvi­ous­ly a high­ly sen­si­tive top­ic and a large num­ber of peo­ple both inside and out­side the US nation­al secu­ri­ty state prob­a­bly don’t want to see major pub­lic scruti­ny of that rela­tion­ship. For exam­ple, it turns out Joe Biden’s cur­rent Direc­tor of Nation­al Intel­li­gence, Avril Haines, was a Palan­tir con­sul­tant from July 5, 2017 to June 23, 2020, plac­ing her at the com­pa­ny dur­ing the peri­od of this new­ly dis­cov­ered Trump admin­is­tra­tion spy­ing.

So what was Haines doing at Palan­tir dur­ing this peri­od? Well, here’s where it starts look­ing bad. Because as the fol­low­ing arti­cle describes, Haines scrubbed her work at Palan­tir short­ly after being select­ed for a poten­tial Biden tran­si­tion team in the sum­mer of 2020. It’s not a great look.

But anoth­er part of the rea­son the selec­tion of Haines as a nation­al secu­ri­ty fig­ure for the Biden admin­is­tra­tion raised the ire of so many on Left was because of the role she played in inves­ti­gat­ing the Bush admin­is­tra­tion’s War on Ter­ror tor­ture inter­ro­ga­tion pro­grams and the Oba­ma admin­is­tra­tion’s drone war­fare pro­grams. The way crit­ics see it, Haines effec­tive­ly pro­tect­ed the CIA was mean­ing­ful reper­cus­sions over the role it played in the tor­ture and nor­mal­ized the drone pro­gram. She also voiced her sup­port for for­mer CIA direc­tor Gina Haspel in 2018 despite the role Haspel played in for­mu­lat­ing those tor­ture pro­grams. Haines’s defend­ers view these as nit­picky crit­i­cisms of some­one who suc­cess­ful­ly reigned in US drone war­fare poli­cies and pressed for max­i­mal dis­clo­sures in the tor­ture report.

So Haines is a rather con­tro­ver­sial fig­ure out­side of her work for Palan­tir. But it’s also not hard to imag­ine why Palan­tir would have been very inter­est­ed in hir­ing her. Haines has the cru­cial expe­ri­ence of legal­ly vet­ting intel­li­gence pro­grams, some­thing that would obvi­ous­ly be an invalu­able skill set for a com­pa­ny like Palan­tir. And that brings us to Haines’s answer as to what it was she was doing at Palan­tir: accord­ing to Haines, she was most­ly just focused on diver­si­ty devel­op­ment and men­tor­ing the careers of the young women work­ing there. That was her role for near­ly three years. Diver­si­ty train­ing.

Sure it’s pos­si­ble Palan­tir hired Haines pri­mar­i­ly for diver­si­ty train­ing for three years and the com­pa­ny just ignored her invalu­able expe­ri­ence vet­ting intel­li­gence pro­grams. But is that a real­is­tic answer? Of course not. It com­plete­ly smacks of being a cov­er up. Now, the fact that Haines does­n’t want to talk about what she actu­al­ly did at Palan­tir does­n’t mean she was involved with a ‘Worse than Water­gate’ Trump admin­is­tra­tion ille­gal domes­tic spy­ing oper­a­tion. But it does sug­gest it’s going to be hard­er than it should be get­ting answers about what role Palan­tir may have played in this lat­est scan­dal:

The Dai­ly Beast

The Proxy War Over a Top Biden Advis­er

Deputy CIA Direc­tor Avril Haines restrict­ed drone strikes and helped thou­sands of refugees, but enraged the left over tor­ture. What does her rise mean for a Biden pres­i­den­cy?

by Spencer Ack­er­man
Sr. Nation­al Secu­ri­ty Cor­re­spon­dent
Updat­ed Jul. 07, 2020 2:51PM ET
Pub­lished Jul. 06, 2020 3:41AM ET

The rise of a for­mer deputy CIA direc­tor on Joe Biden’s tran­si­tion team is draw­ing furi­ous objec­tion from the left—to the shock of her col­leagues in the Oba­ma admin­is­tra­tion, who believe Avril Haines’ record in gov­ern­ment ought to endear her to pro­gres­sives.

In late June, the Biden cam­paign announced that Haines, an attor­ney who served as deputy direc­tor of the CIA from 2013 to 2015, will helm the for­eign pol­i­cy and nation­al secu­ri­ty aspects of a poten­tial Biden tran­si­tion team.

To activists, secu­ri­ty experts, con­gres­sion­al aides who are more left than liberal—as well as main­stream human rights cam­paign­ers and at least one ex-senator—Haines’ ele­va­tion is wor­ri­some or unac­cept­able. She approved an “account­abil­i­ty board” that spared CIA per­son­nel reprisal for spy­ing on the Senate’s tor­ture inves­ti­ga­tors, and was part of the team that redact­ed their land­mark report. After the admin­is­tra­tion end­ed, Haines sup­port­ed Gina Haspel for CIA direc­tor, some­one direct­ly impli­cat­ed in CIA tor­ture, a deci­sion that remains raw amongst pro­gres­sive activists. Until late June, she con­sult­ed for the Trump-favorite data firm Palan­tir, which emerged from the CIA.

“This is a pret­ty omi­nous sig­nal about what is to come” in a Biden admin­is­tra­tion, said a Sen­ate staffer who works on nation­al secu­ri­ty issues. “To have the deputy CIA direc­tor tout­ed for her record in advanc­ing human rights and respect for the rule of law I don’t think can be ade­quate­ly squared with not only her record but her delib­er­ate choic­es of advo­ca­cy.”

To Oba­ma admin­is­tra­tion alum­ni who are more lib­er­al than left, the antipa­thy for Haines is stun­ning. Haines was per­haps the lead­ing voice inside the admin­is­tra­tion for restrict­ing the drone cam­paign. She was “a voice of restraint on all coun­tert­er­ror­ism issues,” said Harold Koh, the for­mer State Depart­ment legal advis­er. As deputy nation­al secu­ri­ty advis­er, she was prin­ci­pal­ly respon­si­ble for increas­ing refugee admis­sions against mas­sive nativist head­winds. Haines, her old col­leagues say, kept press­ing to trans­fer detainees out of Guan­tanamo Bay when oth­ers con­ced­ed defeat.
...

The divide between lib­er­al and left per­cep­tions of Haines high­lights a cross­roads for the future of Demo­c­ra­t­ic nation­al secu­ri­ty pol­i­cy and for a prospec­tive Biden pres­i­den­cy. Lib­er­als tend to view Obama’s main­te­nance of the war on ter­ror, how­ev­er cir­cum­scribed, as unfor­tu­nate but under­stand­able. A rein­vig­o­rat­ed left views it as an epic, dis­cred­it­ing mis­take. Behind its dis­sat­is­fac­tion with Haines is a fear that Biden will restore the Oba­ma lega­cy, rather than expand its hori­zons to, among oth­er things, rolling back the coun­tert­er­ror­ism appa­ra­tus.

...

Haines is not part of Biden’s for­eign-pol­i­cy inner cir­cle, like for­mer Deputy Sec­re­tary of State Tony Blinken; ex-Pen­ta­gon pol­i­cy offi­cial Bri­an McK­eon; Jake Sul­li­van, Biden’s vice-pres­i­den­tial nation­al secu­ri­ty advis­er; and Car­lyn Reichel, his vice-pres­i­den­tial for­eign-pol­i­cy speech­writer. The cam­paign says she won’t have a lead­ing role staffing the admin­is­tra­tion, despite a per­cep­tion to the con­trary on the left. Instead, she’ll con­vert Biden’s cam­paign pledges on those sub­jects into poli­cies for the first year of his pres­i­den­cy. Many expect her to get a senior posi­tion in the admin­is­tra­tion.

“The tran­si­tion oper­a­tion will be focused on respon­si­bly devel­op­ing the readi­ness of a poten­tial new admin­is­tra­tion to serve the Amer­i­can peo­ple,” said Ted Kauf­man, the for­mer sen­a­tor in charge of Biden’s tran­si­tion.

Inter­views with nine for­mer Oba­ma offi­cials, as well as admir­ers in the human rights com­mu­ni­ty, echoed with paeans to Haines’ warmth, dili­gence and com­mit­ment to the law. As State Depart­ment attor­ney dur­ing the late Bush admin­is­tra­tion, Haines unearthed and shep­herd­ed through the Sen­ate 90 lan­guish­ing treaties. “With­out Avril, the Bush admin­is­tra­tion would not have had this very good record on treaties,” said John Bellinger, Koh’s pre­de­ces­sor as State Depart­ment legal advis­er. “I can’t think of a bad thing to say about Avril, I just think she’s a super­star.”

A detail to the Sen­ate For­eign Rela­tions Com­mit­tee put her on the radar of Biden, then the chair­man. By late 2010, she tran­si­tioned to the White House, where she was deputy legal advis­er before ascend­ing to the legal advis­er’s job the next year. There she chaired the inter­a­gency lawyers’ group that would con­vene to con­sid­er the “tar­get­ed killing” enterprise—that is, drone strikes. She came to see it as oper­at­ing with­out mean­ing­ful con­straint and quick­ly part­nered with White House coun­tert­er­ror­ism advis­er John Bren­nan, the CIA vet­er­an.

“We want­ed to make sure that the coun­tert­er­ror­ism pro­gram and any type of lethal strikes that we might take would be very sharply cav­erned with­in a frame­work that made cer­tain stip­u­la­tions [and] cri­te­ria before any strike was tak­en,” said Bren­nan. “We all approached it from our var­i­ous port­fo­lios in a man­ner that lim­it­ed the num­ber of times that strikes would be autho­rized. Avril and I bore the scars of a lot of the push­back that we received from coun­tert­er­ror­ism pro­po­nents that want­ed to have more lat­i­tude in car­ry­ing out strikes.”

Koh, a con­trib­u­tor to the process, remem­bered Haines as a force for curb­ing the drones. “A lot of peo­ple char­ac­ter­ize them­selves as voic­es of restraint, but she real­ly was. ‘That’s ille­gal, we’re not gonna do that,’ she would say. She showed guts,” Koh recalled.

Haines remem­bered press­ing for a process that would ensure drone strikes would occur “only in the rarest cir­cum­stances, when it’s absolute­ly nec­es­sary.”

The result was Obama’s 2013 2013 pol­i­cy-plan­ning guid­ance. It required “near cer­tain­ty” that both some­one tar­get­ed “is in fact the law­ful target”—a stan­dard that did not pre­vi­ous­ly exist—and that civil­ians would not be killed. The drone attacks dimin­ished. In 2010, the high-water mark of the bom­bard­ment, the CIA launched 122 strikes in Pak­istan alone. After the guid­ance was issued, there were 61 Pak­istan strikes in Obama’s entire sec­ond term. But while the drone attacks dimin­ished, human rights groups, as well as rel­a­tives and sur­vivors of drone strikes, dis­put­ed that civil­ians had stopped dying from the lethal activ­i­ties in sig­nif­i­cant num­bers. And while the drones were placed under restric­tions, they per­sist­ed.

Haines sup­port­ed restraint. She did not, she said, sup­port abo­li­tion, which she did not con­sid­er “real­is­tic.” Her for­mer col­leagues say that the only one who could have decid­ed on abo­li­tion was Oba­ma.

“The drone pro­gram exist­ed and wasn’t going away. Pres­i­dent Oba­ma saw the risks of abuse in the pro­gram and tasked Avril with mak­ing it law-abid­ing,” said Pow­er. “Avril sought to put a lethal instru­ment of U.S. pow­er into a legal frame­work, to min­i­mize the risk of civil­ian casu­al­ties, and to give a pro­gram shroud­ed in secre­cy far more trans­paren­cy.”

Andrea Pra­sow, the act­ing Wash­ing­ton direc­tor of Human Rights Watch, has resist­ed the war on ter­ror since its incep­tion. She cred­it­ed Haines with increas­ing trans­paren­cy around the drone strikes—though, in 2016, the Oba­ma admin­is­tra­tion released a civil­ian death tal­ly that human rights groups con­sid­ered a cyn­i­cal under­count—and said Haines did not share the “just-trust-us approach, ‘we’re the good guys’” that she saw from oth­er Oba­ma offi­cials. At the same time, Pra­sow con­tin­ued, “I don’t know how you rec­on­cile the drone pro­gram with any­one who believes in human rights and inter­na­tion­al law.”

Haines said she “under­stood the con­cern expressed by some that the process that was put in place legit­imized the pro­gram, but if you come to the con­clu­sion that the pro­gram will remain in place, hav­ing a rig­or­ous process and a clear, trans­par­ent legal frame­work that pro­motes account­abil­i­ty is crit­i­cal, espe­cial­ly one we can live with as oth­er coun­tries begin to have access to such weapons.”

Bren­nan, Obama’s sec­ond-term CIA direc­tor, took Haines, an out­sider and a lawyer, to Lan­g­ley as his deputy in 2013, “to chal­lenge many con­ven­tion­al wis­dom or think­ing or prac­tices with­in CIA.” Asked what Haines’ lega­cy at the agency was, Bren­nan called her a “tremen­dous men­tor and role mod­el to young offi­cers, espe­cial­ly to women,” as well as aid­ing with a struc­tur­al over­haul unveiled in 2015 and ensur­ing CIA lawyers under­stood Obama’s coun­tert­er­ror­ism restric­tions. A dif­fer­ent for­mer senior CIA offi­cial, how­ev­er, imme­di­ate­ly answered, “she had that shit-burg­er to deal with”—meaning the Sen­ate intel­li­gence committee’s tor­ture inves­ti­ga­tion.

Led by Sen. Dianne Fein­stein, the years-long inves­ti­ga­tion found that the CIA tor­ture was vast­ly more sadis­tic than known; use­less for coun­tert­er­ror­ism; and enveloped in an edi­fice of lies so exten­sive as to con­sti­tute a dis­in­for­ma­tion cam­paign against Con­gress and the pub­lic. The CIA, resist­ing those con­clu­sions, took the fate­ful step of secret­ly access­ing Sen­ate inves­ti­ga­tors’ work prod­uct on a shared pri­vate network—enraging Feinstein—and request­ed the Jus­tice Depart­ment pros­e­cute lead inves­ti­ga­tor Daniel Jones. A CIA inspec­tor gen­er­al report said the agency per­son­nel involved in the spy­ing exhib­it­ed a “lack of can­dor” about the episode.

Haines played two roles over the report. First, she was part of the CIA team, sup­port­ed by the White House, that spent months nego­ti­at­ing with the Sen­ate over how much of the report to declas­si­fy. It infu­ri­at­ed Fein­stein and her allies, who saw the pur­pose of the exer­cise as con­ceal­ing the report’s find­ings. “My rec­ol­lec­tion was that Avril was push­ing as vig­or­ous­ly as she could for min­i­mal redac­tion,” said Denis McDo­nough, Obama’s White House chief of staff. Those rec­ol­lec­tions are not shared by oth­ers in the process, who remem­ber Haines press­ing to obscure the Sen­ate nar­ra­tive. Haines would not com­ment about it for this sto­ry.

“It was not my sense that her goal was to cov­er up tor­ture,” said Human Rights Watch’s Pra­sow. “Was that the out­come? Sure.”

Sec­ond, Bren­nan appoint­ed an “account­abil­i­ty board” to assess the intru­sion. Its find­ings clawed back the CIA inspec­tor general’s assess­ment, found no rea­son to dis­ci­pline those who spied on their Sen­ate over­seers, and crit­i­cized the Sen­ate. Bren­nan recused him­self, leav­ing Haines to accept the board’s con­clu­sions, which she did in one of her last acts before return­ing to the White House in 2015 as deputy nation­al secu­ri­ty advis­er.

“I found the Board’s review and con­clu­sions to be per­sua­sive and con­se­quent­ly, I accept­ed their rec­om­men­da­tions. I have no trou­ble believ­ing that peo­ple dis­agreed with the Board’s con­clu­sions or, for that mat­ter, my accep­tance of them,” Haines said. “Both Sen­ate staff and CIA per­son­nel felt pas­sion­ate about the sit­u­a­tion. Per­son­nel on the agency side felt wronged, like the Sen­ate staff had gone after them, and the peo­ple on the Sen­ate side felt like the agency folks had spied on them. I hon­est­ly think both sides have a mis­im­pres­sion of the oth­er side’s intent and I under­stand that oth­ers will not have come to the same con­clu­sion. But again this has noth­ing to do with the RDI [Ren­di­tion, Deten­tion, and Inter­ro­ga­tion] report or the pro­gram and what I think about tor­ture, which I believe is immoral and unac­cept­able.”

Mark Udall was a Demo­c­ra­t­ic sen­a­tor from Col­orado on the intel­li­gence com­mit­tee when it fin­ished the tor­ture report. Asked about Haines’ role with Biden, Udall said: “If our coun­try is going to turn the page on the dark chap­ter of our his­to­ry that was the CIA’s tor­ture pro­gram, we need to stop nom­i­nat­ing and con­firm­ing indi­vid­u­als who led this ter­ri­ble pro­gram and helped cov­er it up. I trust Joe Biden to ensure his admin­is­tra­tion’s intel­li­gence agen­cies under­stand the griev­ous mis­takes the CIA com­mit­ted through its tor­ture pro­gram and to only nom­i­nate intel­li­gence offi­cials who are ded­i­cat­ed to chang­ing the cul­ture at the CIA.”

Haines returned to the White House for what her col­leagues con­sid­er per­haps her finest hour. The Syr­i­an civ­il war and the so-called Islam­ic State prompt­ed a dire refugee exo­dus. It also prompt­ed a nativist back­lash on both sides of the Atlantic. Repub­li­can gov­er­nors, con­flat­ing ISIS with those flee­ing them, refused to reset­tle refugees. Haines took charge of expand­ing the admis­sions. Ron­nie New­man, a for­mer NSC offi­cial, remem­bered Haines lever­ag­ing her CIA pedi­gree against intran­si­gent secu­ri­ty agen­cies. “She was able to say con­vinc­ing­ly and per­sua­sive­ly not only we could live up to our human­i­tar­i­an com­mit­ments but also keep the nation safe,” New­man said. “There were life and death con­se­quences for peo­ple and that was what was great about work­ing with her. Every refugee count­ed.”

Haines’ work raised refugee admis­sions from 70,000 to 85,000. As the Oba­ma admin­is­tra­tion wound down in fall 2016, she got the admis­sions totals raised again, to 110,000 for fis­cal 2017. In a speech to Human Rights First, Haines framed embrac­ing refugees as a coun­tert­er­ror­ism mea­sure, since “when we sup­port and care for refugees, we con­tra­dict [extrem­ists’] mes­sage.” Ben Rhodes, one of Obama’s chief for­eign-pol­i­cy aides, reflect­ed, “Not a sin­gle human being besides Barack Oba­ma did more than Avril to get more refugees into this coun­try.”

A record like that stunned Democ­rats when, in 2018, Haines joined a cho­rus of for­mer intel­li­gence offi­cials sup­port­ing Gina Haspel for CIA direc­tor. To anti-tor­ture activists, it was nau­se­at­ing to per­mit some­one who played a lead­ing role in tor­ture to run the CIA—and the inevitable con­se­quence of sup­press­ing the Sen­ate report. As Demo­c­ra­t­ic oppo­si­tion to her nom­i­na­tion crest­ed, the White House crowed over the ex-intel­li­gence offi­cials’ sup­port.

For many on the left, this moment defines Haines. “Even in the Trump era, with the sup­posed ‘#Resis­tance’ ral­ly­ing cry of con­gres­sion­al Democ­rats, that [Haines] went on the record and endorsed Haspel speaks to the depths of a com­mit­ment to a sim­i­lar­ly law­less enter­prise,” said the Sen­ate staffer. Added the leader of a pro­gres­sive non­prof­it that works on nation­al secu­ri­ty issues who request­ed anonymi­ty out of con­cern for pro­fes­sion­al reprisal, “Being where any decent per­son should be on a few issues doesn’t can­cel out an endorse­ment of tor­tur­ers.” Danielle Bri­an, the exec­u­tive direc­tor of the Project on Gov­ern­ment Over­sight, said, “Her sup­port was used by the admin­is­tra­tion to legit­imize the nom­i­na­tion, and that’s a black mark against Haines’ record.”

Haines would not com­ment for this sto­ry about her sup­port of Haspel. Sources famil­iar with her think­ing on the episode said that she had come to see Haspel as an inde­pen­dent voice with­in CIA and thought that of any­one Trump would nom­i­nate to the posi­tion, Haspel would be best equipped to push back against inap­pro­pri­ate uses of the agency. In an April co-authored piece for For­eign Pol­i­cy, Haines observed that Trump’s long-pre­dict­ed politi­ciza­tion of the intel­li­gence agen­cies has man­i­fest­ed, some­thing many of Haspel sup­port­ers backed her to pre­vent.

After the Oba­ma admin­is­tra­tion end­ed, Haines took sev­er­al aca­d­e­m­ic and con­sult­ing posi­tions. One of them was with Palan­tir, the data firm allied with Trump that, among oth­er things, aid­ed ICE in round­ing up undoc­u­ment­ed immi­grants. Accord­ing to Palan­tir, Haines con­sult­ed on pro­mot­ing diver­si­ty with­in the company’s hir­ing from July 5, 2017 to June 23, short­ly after her posi­tion with the Biden tran­si­tion was announced. As The Inter­cept first report­ed, Palan­tir quick­ly dis­ap­peared from her Brook­ings Insti­tu­tion biog­ra­phy, smack­ing of a white­wash. Brook­ings told The Dai­ly Beast that Haines’ office had request­ed an update scrubbed of non-active affil­i­a­tions broad­er than Palan­tir. A Biden tran­si­tion offi­cial said Haines removed sev­er­al affil­i­a­tions from her bio, not just Palan­tir, after end­ing those affil­i­a­tions as part of her onboard­ing to the tran­si­tion.

“The vast major­i­ty of my work for Palan­tir was relat­ed to diver­si­ty and inclu­sion, with a par­tic­u­lar focus on gen­der. For the most part, this involved vis­it­ing with dif­fer­ent offices, talk­ing to those in the work­force about their expe­ri­ences, occa­sion­al­ly men­tor­ing some of the remark­able young women who work there and sug­gest­ing ways in which they might pro­mote diver­si­ty and inclu­sion,” she said. “This is an issue I feel pas­sion­ate about and on which we need to do bet­ter not just at the CIA but across the nation­al secu­ri­ty work­force in gov­ern­ment.”

Haines’ left crit­ics con­sid­er her Palan­tir work egre­gious, swampy and a cap­stone for her career. “It’s inter­est­ing to do diver­si­ty for a com­pa­ny found­ed by Peter Thiel,” who has mused that wom­ens’ polit­i­cal empow­er­ment is a neg­a­tive for “cap­i­tal­ist democ­ra­cy,” observed Jeff Hauser of the Cen­ter for Eco­nom­ic and Pol­i­cy Research.

“Those who engage in revolv­ing-door for-prof­it nation­al secu­ri­ty firms like Palan­tir, there’s some­thing redo­lent of the cor­rup­tion the pro­gres­sive left is fight­ing against,” said the Sen­ate staffer. “We should absolute­ly be able to expect that a Demo­c­ra­t­ic nation­al secu­ri­ty leader will both be humane as relates to refugees—and also not cov­er up for tor­ture, pro­mote tor­tur­ers, and take pay­checks from some of the world’s most malev­o­lent cor­po­ra­tions,” the non­prof­it leader added.

More broad­ly, the con­cerns with Haines on the left under­score an exhaus­tion over Obama’s cau­tious embrace of the war on ter­ror and a fear that Biden will con­tin­ue it. Con­strain­ing the war on ter­ror instead of dis­man­tling it did noth­ing to con­front the post‑9/11 nativist secu­ri­ty para­noia that Trump rode to pow­er. Once he did, all the work Haines did to cir­cum­scribe the drone strikes van­ished as Trump inten­si­fied the bom­bard­ment and returned it to the shad­ows. The 110,000-refugee ceil­ing she raised crashed to its foun­da­tions. All that remains is the war on ter­ror.

“When we look at the con­ti­nu­ities between Bush and Oba­ma, we should be con­cerned that we’re going to return in a Biden admin­is­tra­tion to a kind of sta­tus quo. A return to trans­paren­cy and legal­i­ty is hard­ly enough,” observed Nikhil Pal Singh of New York Uni­ver­si­ty, author of Race and America’s Long War. “That’s just hand­ing the baton back and forth between two types of approach­es that are deeply flawed, unjust and pro­vide no durable secu­ri­ty frame­work.”

The left­most Oba­ma alum­ni want the Biden team to lis­ten to the dis­sat­is­fac­tion and trans­late it into pol­i­cy. “The main take­away from this con­tro­ver­sy is that the Biden cam­paign ought to reach out to pro­gres­sives and hear them out on mat­ters of for­eign pol­i­cy as much as it does on domes­tic social and eco­nom­ic mat­ters,” said Rob Mal­ley, pres­i­dent of the Inter­na­tion­al Cri­sis Group and Haines’ for­mer col­league on the NSC.

Haines said that’s what she wants as well.

“Yes, I’m absolute­ly open to it. There’s no ques­tion. What the Bush admin­is­tra­tion called the glob­al war on ter­ror and what the Oba­ma Admin­is­tra­tion called the con­flict with al Qae­da and asso­ci­at­ed forces, can­not sim­ply exist for­ev­er on auto­mat­ic,” she told The Dai­ly Beast. “To the extent the con­cern would be ‘Is she some­body who rep­re­sents just a return to the poli­cies of the Oba­ma admin­is­tra­tion, sim­ply pro­mot­ing con­straint but not actu­al­ly chang­ing the land­scape,’ that’s not a con­cern with me. We have to rethink things.”
Spencer Ack­er­man

———–

“The Proxy War Over a Top Biden Advis­er” by Spencer Ack­er­man; The Dai­ly Beast; 07/06/2020

“After the Oba­ma admin­is­tra­tion end­ed, Haines took sev­er­al aca­d­e­m­ic and con­sult­ing posi­tions. One of them was with Palan­tir, the data firm allied with Trump that, among oth­er things, aid­ed ICE in round­ing up undoc­u­ment­ed immi­grants. Accord­ing to Palan­tir, Haines con­sult­ed on pro­mot­ing diver­si­ty with­in the company’s hir­ing from July 5, 2017 to June 23, short­ly after her posi­tion with the Biden tran­si­tion was announced. As The Inter­cept first report­ed, Palan­tir quick­ly dis­ap­peared from her Brook­ings Insti­tu­tion biog­ra­phy, smack­ing of a white­wash. Brook­ings told The Dai­ly Beast that Haines’ office had request­ed an update scrubbed of non-active affil­i­a­tions broad­er than Palan­tir. A Biden tran­si­tion offi­cial said Haines removed sev­er­al affil­i­a­tions from her bio, not just Palan­tir, after end­ing those affil­i­a­tions as part of her onboard­ing to the tran­si­tion.”

All of a sud­den her three years of work at Palan­tir dis­ap­peared from her Brook­ings Insti­tu­tion biog­ra­phy. It’s not hard to imag­ine rea­sons for this. Palan­tir is a scan­dalous com­pa­ny, espe­cial­ly for a puta­tive Demo­c­ra­t­ic admin­is­tra­tion, with or with­out a spy­ing scan­dal. But it’s also not hard to imag­ine that the work Haines actu­al­ly did for Palan­tir is the kind of work she real­ly does­n’t want to talk about, which is why her claims of focus­ing on diver­si­ty and inclu­sion ring to hol­low. Why scrub your diver­si­ty and inclu­sion work?

...
“The vast major­i­ty of my work for Palan­tir was relat­ed to diver­si­ty and inclu­sion, with a par­tic­u­lar focus on gen­der. For the most part, this involved vis­it­ing with dif­fer­ent offices, talk­ing to those in the work­force about their expe­ri­ences, occa­sion­al­ly men­tor­ing some of the remark­able young women who work there and sug­gest­ing ways in which they might pro­mote diver­si­ty and inclu­sion,” she said. “This is an issue I feel pas­sion­ate about and on which we need to do bet­ter not just at the CIA but across the nation­al secu­ri­ty work­force in gov­ern­ment.”

Haines’ left crit­ics con­sid­er her Palan­tir work egre­gious, swampy and a cap­stone for her career. “It’s inter­est­ing to do diver­si­ty for a com­pa­ny found­ed by Peter Thiel,” who has mused that wom­ens’ polit­i­cal empow­er­ment is a neg­a­tive for “cap­i­tal­ist democ­ra­cy,” observed Jeff Hauser of the Cen­ter for Eco­nom­ic and Pol­i­cy Research.

“Those who engage in revolv­ing-door for-prof­it nation­al secu­ri­ty firms like Palan­tir, there’s some­thing redo­lent of the cor­rup­tion the pro­gres­sive left is fight­ing against,” said the Sen­ate staffer. “We should absolute­ly be able to expect that a Demo­c­ra­t­ic nation­al secu­ri­ty leader will both be humane as relates to refugees—and also not cov­er up for tor­ture, pro­mote tor­tur­ers, and take pay­checks from some of the world’s most malev­o­lent cor­po­ra­tions,” the non­prof­it leader added.
...

We’ll see if any ques­tions about poten­tial roles Palan­tir may have played in the Trump admin­is­tra­tion’s domes­tic spy­ing activ­i­ties actu­al­ly end up get­ting asked. It’s unlike­ly. But if those ques­tions do end up get­ting asked it will be inter­est­ing to learn more about the diver­si­ty and inclu­sion train­ing being done at one of the world’s lead­ing fas­cist-owned Big Data NSA-for-hire ser­vice providers.

This arti­cle talks about how the US soft­ware devel­op­er for US Inteligence (fund­ed by Peter Thiele) Palan­tir signed secre­tive con­tracts with the Greeks and had secre­tive talks with EU Pres­i­dent (orig­i­nal­ly from Ger­many) Ursu­la von der Leyen as well as with the then EU’s com­pe­ti­tion com­mis­sion­er, Mar­grethe Vestager, who is now in charge of mak­ing the EU fit for the dig­i­tal age. The arti­cle rais­es con­cerns of vio­la­tion of EU’s data pro­tec­tion laws includ­ing Palentir’s access to Europol data and inves­ti­ga­tions and wit­ness tes­ti­mo­ny.

Palentir’s soft­ware “Gotham” has been used by intel­li­gence ser­vices in the UK, the Nether­lands, Den­mark and France and was built for inves­tiga­tive analy­sis. Some Palan­tir engi­neers call what it does “nee­dle-in-haystack” analy­sis that agen­cies can use to look for bad actors hid­ing in com­plex net­works.

Their soft­ware also claims to be pre­dic­tive of crime but the accu­ra­cy of that is con­tro­ver­sial and has not been dis­closed. There was a con­cern that there is an imbal­ance of pow­er with knowl­edge of data use and between soft­ware firms and the pub­lic inter­est. Pri­vate pow­er over pub­lic process­es is grow­ing expo­nen­tial­ly with access to data and tal­ent.

Palen­tir is also get­ting into the ground floor of a new cloud soft­ware inter­face require­ments for the EU called of GAIA‑X.

Implic­it­ly if you read between the lines of the arti­cle, Palen­tar is a soft­ware that is mar­ket­ed for intel­li­gence gath­er­ing but is like­ly an espi­onage tool used to acquire data on indi­vid­u­als to be used for polit­i­cal manip­u­la­tion.

Impo­trant con­nec­tions to note are Palentir’s, CEO Alex Karp, stud­ied in Ger­many at Frank­furt Uni­ver­si­ty under the influ­en­tial philoso­pher Jür­gen Haber­mas. Michael Krat­sios was chief tech­nol­o­gy advis­er to then-pres­i­dent, Don­ald Trump. Krat­sios joined the White House from a role as chief of staff to Peter Thiel, the bil­lion­aire Sil­i­con Val­ley tech investor and founder of Palan­tir, key investor in Face­book, and Pay­pal.

The Guardian, April 2, 2011
See­ing stones: pan­dem­ic reveals Palan­tir’s trou­bling reach in Europe Covid has giv­en Peter Thiel’s secre­tive US tech com­pa­ny new oppor­tu­ni­ties to oper­ate in Europe in ways some cam­paign­ers find wor­ry­ing

by Daniel How­den, Apos­tol­is­Fo­tiadis, Ludek Stavi­no­ha, Ben Holst.

https://www.theguardian.com/world/2021/apr/02/seeing-stones-pandemic-reveals-palantirs-troubling-reach-in-europe?CMP=Share_iOSApp_Other

See­ing stones: pan­dem­ic reveals Palan­tir’s trou­bling reach in Europe
Covid has giv­en Peter Thiel’s secre­tive US tech com­pa­ny new oppor­tu­ni­ties to oper­ate in Europe in ways some cam­paign­ers find wor­ry­ing

The 24 March, 2020 will be remem­bered by some for the news that Prince Charles test­ed pos­i­tive for Covid and was iso­lat­ing in Scot­land. In Athens it was mem­o­rable as the day the traf­fic went silent. Twen­ty-four hours into a hard lock­down, Greeks were accli­ma­tis­ing to a new real­i­ty in which they had to send an SMS to the gov­ern­ment in order to leave the house. As well as mil­lions of text mes­sages, the Greek gov­ern­ment faced extra­or­di­nary dilem­mas. The Euro­pean Union’s most vul­ner­a­ble econ­o­my, its old­est pop­u­la­tion along with Italy, and one of its weak­est health sys­tems faced the first wave of a pan­dem­ic that over­whelmed rich­er coun­tries with few­er pen­sion­ers and stronger health pro­vi­sion. The car­nage in Italy loomed large across the Adri­at­ic.

One Greek who did go into the office that day was Kyr­i­akos Pier­rakakis, the min­is­ter for dig­i­tal trans­for­ma­tion, whose sig­na­ture was inked in blue on an agree­ment with the US tech­nol­o­gy com­pa­ny, Palan­tir. The deal, which would not be revealed to the pub­lic for anoth­er nine months, gave one of the world’s most con­tro­ver­sial tech com­pa­nies access to vast amounts of per­son­al data while offer­ing its soft­ware to help Greece weath­er the Covid storm. The zero-cost agree­ment was not reg­is­tered on the pub­lic pro­cure­ment sys­tem, nei­ther did the Greek gov­ern­ment car­ry out a data impact assess­ment – the man­dat­ed check to see whether an agree­ment might vio­late pri­va­cy laws.

The ques­tions that emerge in pan­dem­ic Greece echo those from across Europe dur­ing Covid and show Palan­tir extend­ing into sec­tors from health to polic­ing, avi­a­tion to com­merce and even acad­e­mia. A months-long joint inves­ti­ga­tion by the Guardian, Light­house Reports and Der Spiegel used free­dom of infor­ma­tion laws, offi­cial cor­re­spon­dence, con­fi­den­tial sources and report­ing in mul­ti­ple coun­tries to piece togeth­er the Euro­pean activ­i­ties of one of the most secre­tive com­pa­nies in the world. The find­ings raise seri­ous ques­tions over the way pub­lic agen­cies work with Palan­tir and whether its soft­ware can work with­in the bounds of Euro­pean laws in the sen­si­tive areas where it is being used, or per­form in the way the com­pa­ny promis­es.

Greece was not the only coun­try tempt­ed by a Covid-relat­ed free tri­al. Palan­tir was already embed­ded in the NHS, where a no-bid con­tract val­ued at £1 was only revealed after data pri­va­cy cam­paign­ers threat­ened to take the UK gov­ern­ment to court. When that tri­al peri­od was over the cost of con­tin­u­ing with Palan­tir came in at £24m.

The com­pa­ny has also been con­tract­ed as part of the Nether­lands’ Covid response and pitched at least four oth­er Euro­pean coun­tries, as well as a clutch of EU agen­cies. The Palan­tir one-pager that Germany’s health min­istry released after a free­dom of infor­ma­tion request described Europe as the company’s “focus of activ­i­ties”.

Found­ed in Cal­i­for­nia in 2003, Palan­tir may not have been cold-call­ing around Euro­pean gov­ern­ments. It has, at times, had a unique­ly pow­er­ful busi­ness devel­op­ment ally in the form of the US gov­ern­ment.
On 23 March, the EU’s Cen­tre for Dis­ease Con­trol (ECDC) received an email from their coun­ter­parts at the US CDC, extolling their work with Palan­tir and say­ing the com­pa­ny had asked for an intro­duc­tion.

Palan­tir said it was nor­mal prac­tice for some of its “gov­ern­ment cus­tomers to serve as ref­er­ence for oth­er prospec­tive cus­tomers”. It said the ECDC turned down its invi­ta­tion “out of con­cern of a risk of the con­tact being per­ceived as prej­u­dic­ing ECDC’s inde­pen­dence”.

PHOTO CAPTION: A Palan­tir ban­ner out­side the New York Stock Exchange on the day of its ini­tial pub­lic offer­ing on 30 Sep­tem­ber, 2020. Pho­to­graph: Andrew Kelly/Reuters

The Greek gov­ern­ment has declined to say how it was intro­duced to Palan­tir. But there were senior-lev­el links between Palan­tir, the Trump admin­is­tra­tion and the Greek gov­ern­ment. The US ambas­sador to Greece, Geof­frey Pyatt, has spo­ken pub­licly of the con­tacts between Pier­rakakis and Michael Krat­sios, a Greek-Amer­i­can and chief tech­nol­o­gy advis­er to then-pres­i­dent, Don­ald Trump. Krat­sios joined the White House from a role as chief of staff to Peter Thiel, the bil­lion­aire Sil­i­con Val­ley tech investor and founder of Palan­tir.

When news of Greece’s rela­tion­ship with Palan­tir was dis­closed, it was not by gov­ern­ment offi­cials or local media but by ambas­sador Pyatt. A tele­con­fer­ence fol­lowed in Decem­ber between Greece’s prime min­is­ter, Kyr­i­akos Mit­so­takis, and Palan­tir CEO Alex Karp, where the lat­ter spoke of “deep­en­ing coop­er­a­tion” between them.

Jour­nal­ists who asked for a copy of the agree­ment were refused and it took oppo­si­tion MPs to force dis­clo­sure via par­lia­ment. The tone then abrupt­ly changed.

Eleft­he­rios Che­lioudakis, a data pro­tec­tion lawyer and mem­ber of dig­i­tal rights group Homo Dig­i­tal­is, was among the first peo­ple to read the two-page doc­u­ment and was stunned by what he found. It appeared to give Palan­tir phe­nom­e­nal access to data of exact­ly the scale and sen­si­tiv­i­ty that would seem to require an impact assess­ment. Worse, a revi­sion of the agree­ment one week after the first delet­ed any ref­er­ence to the need to “pseu­do­nymise” the data – to pre­vent it being relat­able to spe­cif­ic indi­vid­u­als. This appears to be in breach of the Gen­er­al Data Pro­tec­tion Reg­u­la­tion (GDPR), the EU law in place since 2018 that gov­erns how the per­son­al infor­ma­tion of peo­ple liv­ing in the EU can be col­lect­ed and processed. Palan­tir says that, to its knowl­edge, pro­cess­ing was lim­it­ed to “open-source pan­dem­ic and high-lev­el Greek state-owned demo­graph­ic data direct­ly rel­e­vant to man­ag­ing the Covid-19 cri­sis”.

PHOTO CAPTION: The Greek prime min­is­ter, Kyr­i­akos Mit­so­takis (cen­tre), and the min­is­ter of dig­i­tal gov­er­nance, Kyr­i­akos Pier­rakakis (left), chat with the US ambas­sador to Greece, Geof­frey Pyatt (right), in Thes­sa­loni­ki, Greece, in Sep­tem­ber 2019. Pho­to­graph: Kostas Tsironis/EPA

The Greek gov­ern­ment has denied shar­ing patient data with Palan­tir, claim­ing that the soft­ware was used to give the prime min­is­ter a dash­board sum­maris­ing key data dur­ing the pan­dem­ic. How­ev­er, the con­tract, seen by the Guardian, specif­i­cal­ly refers to cat­e­gories of data that can be processed and includes per­son­al data. It also includes a clause that has come to be known as an “improve­ment clause”. These claus­es, iden­ti­fied in the rare exam­ples of Palan­tir con­tracts released in answer to free­dom of infor­ma­tion requests, have been stud­ied by Pri­va­cy Inter­na­tion­al, a pri­va­cy watch­dog in the UK. “The improve­ment claus­es in Palantir’s con­tracts, togeth­er with the lack of trans­paren­cy, are con­cern­ing because it enables Palan­tir to improve its prod­ucts based on its cus­tomers’ use of the Palan­tir prod­ucts,” said Pri­va­cy International’s Caitlin Bish­op.

The com­pa­ny rejects this read­ing of their activ­i­ties and states: “Palan­tir does not train algo­rithms on cus­tomer data for Palantir’s own ben­e­fit or to com­mer­cialise and sell to Palantir’s oth­er cus­tomers.”

“We do not col­lect, mine, or sell per­son­al data from or for our cus­tomers,” it said, adding: “Palan­tir does not use its cus­tomers’ data to build, deploy, trans­fer, resell, or repur­pose machine learn­ing or arti­fi­cial intel­li­gence mod­els or ‘algo­rithms’ to oth­er cus­tomers.”

Greece’s data pro­tec­tion author­i­ty has since launched an inves­ti­ga­tion. The gov­ern­ment says it has end­ed coop­er­a­tion with Palan­tir and that all data has been delet­ed.

Lord of the Rings mys­tique

Even by the stan­dards of Sil­i­con Val­ley tech com­pa­nies, Palan­tir has been an out­lier in cre­at­ing a mythol­o­gy around itself. The name is tak­en from the pow­er­ful and per­ilous “see­ing stones” in Tolkien’s Lord of the Rings. Its lead­er­ship often claims the man­tle of defend­ers of the west­ern realm. Ear­ly employ­ees cast them­selves as brave hob­bits and one of Thiel’s co-founders wrote about his depar­ture from the com­pa­ny in a post enti­tled “leav­ing the Shire”.

But Palan­tir polarised opin­ion in the US before the back­lash against big tech. Its crit­ics do not focus on the for­tune its founder Thiel made with Pay­Pal or as an ear­ly investor in Face­book but on his sup­port for Trump. Palan­tir has faced protests in the US over its role in facil­i­tat­ing the Trump administration’s mass depor­ta­tion of undoc­u­ment­ed migrants through its con­tract with US immi­gra­tion enforce­ment agency ICE.

Palan­tir was also report­ed to have been involved in dis­cus­sions over a cam­paign of dis­in­for­ma­tion and cyber­at­tacks direct­ed against Wik­iLeaks and jour­nal­ists such as Glenn Green­wald. It lat­er insist­ed that the project was nev­er put into effect and said its asso­ci­a­tion with smear tac­tics had “served as a teach­able moment”.

And Palan­tir was will­ing to step in at the Pen­ta­gon after Google employ­ees rebelled over its involve­ment in Project Maven, which seeks to use AI in bat­tle­field tar­get­ing.

Until Palan­tir under­took a pub­lic list­ing in Sep­tem­ber last year, rel­a­tive­ly lit­tle was known about its client list beyond ser­vices to the US mil­i­tary, bor­der enforce­ment and intel­li­gence agen­cies.

Media cov­er­age of Palan­tir has been shaped by its unusu­al pro­tag­o­nists as well as its nation­al secu­ri­ty clients. The company’s CEO is Alex Karp, who stud­ied in Ger­many at Frank­furt Uni­ver­si­ty under the influ­en­tial philoso­pher Jür­gen Haber­mas, and often makes cor­po­rate announce­ments in philo­soph­i­cal lan­guage in uncon­ven­tion­al cloth­ing or loca­tions. His most recent mes­sage was tweet­ed from a snowy for­est.

PHOTO CAPTION: Palantir’s CEO Alex Karp. Pho­to­graph: Thibault Camus/AP

Rumours over Palantir’s pos­si­ble involve­ment [with the CIA] in the oper­a­tion to find Osama bin Laden have been met with coy non-denials.
The colour­ful back­sto­ry has added mys­tique to a com­pa­ny which, when it list­ed on the New York stock exchange, had only 125 cus­tomers.

Why did Palan­tir meet Von Der Leyen?

Sophie in ’t Veld, a Dutch MEP, has tracked Palantir’s lob­by­ing of Europe’s cen­tres of pow­er. She notes the company’s unusu­al “prox­im­i­ty to pow­er” and ques­tions how it was that an EU del­e­ga­tion to Wash­ing­ton in 2019 met with US gov­ern­ment offi­cials and only one pri­vate com­pa­ny, Palan­tir. What was dis­cussed, she want­ed to know, when Karp met the pres­i­dent of the Euro­pean com­mis­sion, Ursu­la von der Leyen or when Palan­tir met the then EU’s com­pe­ti­tion com­mis­sion­er, Mar­grethe Vestager, who is now in charge of mak­ing the EU fit for the dig­i­tal age?

PHOTO CAPTION: EU com­mis­sion pres­i­dent Ursu­la Von der Leyen (left) and exec­u­tive vice-pres­i­dent of the Euro­pean Com­mis­sion for A Europe Fit for the Dig­i­tal Age, Mar­grethe Vestager (right), in Brus­sels, Bel­gium, on 19 Feb­ru­ary 2020. Pho­to­graph: Olivi­er Hoslet/EPA

In June 2020, In ‘t Veld sent detailed ques­tions to the com­mis­sion and pub­lished her con­cerns in a blog­post head­lined: “Palan­tir is not our friend”. The com­mis­sion took eight months to give even par­tial answers but the com­pa­ny emailed In ‘t Veld three days after she went pub­lic with her ques­tions, offer­ing a meet­ing. She talked to them but ques­tions why the com­pa­ny felt the need to con­tact “an obnox­ious MEP” to reas­sure her.

In ‘t Veld char­ac­teris­es the commission’s even­tu­al answers as “eva­sive” with offi­cials say­ing no min­utes were kept of the con­ver­sa­tion between Von Der Leyen and Karp because it was on the side­lines of the World Eco­nom­ic Forum at Davos and they already knew each oth­er.

PHOTO CAPTION: Mem­ber of Euro­pean Par­lia­ment, Sophie in ‘t Veld, at Euro­pean par­lia­ment head­quar­ters in Brus­sels, Bel­gium. Pho­to­graph: Wik­tor Dąbkowski/ZUMA Press/Alamy

“There’s some­thing that doesn’t add up here between the cir­cum­vent­ing of pro­cure­ment prac­tices, meet­ings at the high­est lev­el of gov­ern­ment,” said In ‘t Veld, “there’s a lot more beneath the sur­face than a sim­ple soft­ware com­pa­ny.”

For its part, Palan­tir says it is “not a data com­pa­ny” and all data it inter­acts with is “col­lect­ed, owned, and con­trolled by the cus­tomers them­selves, not by Palan­tir.” The com­pa­ny says “it is essen­tial to pre­serve fun­da­men­tal prin­ci­ples of pri­va­cy and civ­il lib­er­ties while using data” and that Palan­tir does not build algo­rithms off its cus­tomers’ data in any form but pro­vides soft­ware plat­forms that serve as the cen­tral oper­at­ing sys­tems for a wide vari­ety of pub­lic and pri­vate sec­tor insti­tu­tions.

Palan­tir said: “We build soft­ware prod­ucts to help our cus­tomers inte­grate and under­stand their own data, but we don’t col­lect, hold, mine, or mon­e­tize data on our own. Of course, our engi­neers may be required to inter­act with some cus­tomer data when they are at cus­tomer sites, but we are not in the busi­ness of col­lect­ing, main­tain­ing, or sell­ing data.”

Europol entan­gle­ment

Covid has been the occa­sion for a new busi­ness dri­ve but Palan­tir did not arrive in Europe with the pan­dem­ic. It has also found oppor­tu­ni­ties in Euro­pean fear of ter­ror­ism and its sense of tech­no­log­i­cal infe­ri­or­i­ty to Sil­i­con Val­ley.

When health con­cerns are dri­ving busi­ness, the soft­ware prod­uct Palan­tir sells is Foundry; when ter­ror­ism fears are open­ing up bud­gets, it is Gotham.

Foundry is built to meet the needs of com­mer­cial clients. One of its cham­pi­ons in Europe is Air­bus, which says the sys­tem has helped iden­ti­fy sup­ply chain effi­cien­cies. Foundry has more recent­ly found its way into gov­ern­ments, and Palantir’s CEO, Karp, has called Foundry an “oper­at­ing sys­tem for gov­ern­ments”.

Gotham has long been used by intel­li­gence ser­vices in the UK, the Nether­lands, Den­mark and France and was built for inves­tiga­tive analy­sis. Some Palan­tir engi­neers call what it does “nee­dle-in-haystack” analy­sis that agen­cies can use to look for bad actors hid­ing in com­plex net­works.

Since 2013 Palan­tir has made a sus­tained dri­ve to embed itself via Gotham in Europe’s police sys­tems.

The first major oppor­tu­ni­ty to do this came at the EU’s law enforce­ment agency, Europol, when it won a ten­der to cre­ate a sys­tem to store and crunch the reams of data from mem­ber states’ police forces. The Europol Analy­sis Sys­tem was meant both to store mil­lions of items of infor­ma­tion – from crim­i­nal records, to wit­ness state­ments to police reports – and crunch this data into action­able intel­li­gence.

The agree­ment signed in Decem­ber 2012 with the French multi­na­tion­al Capgem­i­ni, sub­con­tract­ed the work to Palan­tir and Gotham.

Over the next three years, heav­i­ly redact­ed Europol doc­u­ments, obtained under free­dom of infor­ma­tion laws, tell a sto­ry of repeat­ed delays, “low deliv­ery qual­i­ty” and “per­for­mance issues” relat­ed to Gotham. Amid the blacked-out lines there is men­tion of tech­ni­cal short­com­ings such as the “inabil­i­ty to prop­er­ly visu­al­ize large datasets”.

By May 2016 the issues were so entrenched that Europol agreed a set­tle­ment with Palan­tir, the terms of which they have refused to dis­close. Capgem­i­ni, the con­trac­tor which brought in Palan­tir, also declined to com­ment.

It is also clear that Europol con­sid­ered suing Palan­tir and Capgem­i­ni. In an inter­nal brief­ing doc­u­ment ahead of an Octo­ber 2018 meet­ing of the organisation’s man­age­ment board, it is made clear that lit­i­ga­tion was con­sid­ered but reject­ed: “despite the per­for­mance issues iden­ti­fied [lit­i­ga­tion] is like­ly to lead to cost­ly court pro­ceed­ings for which the out­come is uncer­tain.”

Palan­tir declined to com­ment on these issues specif­i­cal­ly but said: “Any issues aris­ing at Europol had noth­ing to do with the software’s abil­i­ty to meet GDPR or data pro­tec­tion require­ments, and were sole­ly the result of a large, com­plex soft­ware imple­men­ta­tion with mul­ti­ple stake­hold­ers.”
The cau­tion was well advised. Palan­tir has form for suing large pub­lic bod­ies, includ­ing the US army, and win­ning.

When access was request­ed from Europol to all records relat­ing to con­trac­tu­al mat­ters with Palan­tir, 69 doc­u­ments were iden­ti­fied, but the EU agency twice refused full access to 67 on the grounds of “pub­lic secu­ri­ty”. An appeal has been lodged with the Euro­pean ombudsman’s office, a com­plaint that was ruled admis­si­ble and a deci­sion is pend­ing.
The set­tle­ment did not dis­en­tan­gle Europol but it brought the project in-house and the effort to use Gotham as a data repos­i­to­ry was aban­doned but it remained as the main analy­sis com­po­nent. In July 2017, a real-world tri­al of the sys­tem on counter-ter­ror­ism work found Gotham “suf­fer­ing from sig­nif­i­cant per­for­mance issues”. Palan­tir said: “Any issues aris­ing at Europol had noth­ing to do with the software’s abil­i­ty to meet GDPR or data pro­tec­tion require­ments, and were sole­ly the result of a large, com­plex soft­ware imple­men­ta­tion with mul­ti­ple stake­hold­ers.”

Despite these issues, Palan­tir has received €4m (£3.4m) from Europol.
The con­cerns went beyond per­for­mance when the EU’s pri­va­cy watch­dog, the Euro­pean data pro­tec­tion super­vi­sor, began inspec­tions. Heav­i­ly redact­ed copies of their reports in 2018 and 2019 reg­is­ter the inspec­tors’ con­cern that Gotham was not designed to ensure that the Europol ana­lysts made it clear how people’s data had come to be entered into the sys­tem. The absence of this “per­son­al impli­ca­tion” meant the sys­tem could not be guar­an­teed to dis­tin­guish whether some­one was a vic­tim, wit­ness, infor­mant or sus­pect in a crime. This rais­es the prospect of peo­ple being false­ly impli­cat­ed in crim­i­nal inves­ti­ga­tions or, at the very least, that their data may not have been han­dled in com­pli­ance with data pro­tec­tion laws.
Europol, as the data con­troller, said that such data was “treat­ed with the great­est care”.

PHOTO CAPTION: The Europol build­ing in The Hague, Nether­lands. Pho­to­graph: Eva Plevier/Reuters

‘The hottest shit ever in polic­ing’
In 2005, 15 Euro­pean coun­tries signed a deal to boost counter-ter­ror efforts by exchang­ing DNA, fin­ger­prints and vehi­cle reg­is­tra­tion data. This led to an IT buy­ing spree as police author­i­ties sought ways to get their sys­tems to talk to each oth­er. Nor­way was a late­com­er when it signed up in 2009 but in 2016 a high-rank­ing del­e­ga­tion from the Nor­we­gian police flew to Sil­i­con Val­ley to meet Palan­tir. When they returned the force decid­ed to set up a more far-reach­ing sys­tem to be called Omnia, run­ning on Gotham.

The abrupt deci­sion caught the atten­tion of Ole Mar­tin Mortvedt, a for­mer senior police offi­cer near­ing retire­ment who was edit­ing the nation­al police union’s in-house mag­a­zine. When he start­ed ask­ing ques­tions he found it impos­si­ble to estab­lish who had gone to Sil­i­con Val­ley and why the project had been expand­ed. The only rep­re­sen­ta­tive of Palan­tir whom he could talk to in Nor­way was a rel­a­tive­ly junior lawyer.

A frus­trat­ed Mortvedt start­ed call­ing his for­mer pupils from the police acad­e­my where he taught for many years who were now in mid-rank­ing posi­tions in the police. Over the next three years, his police sources described a litany of missed dead­lines.

“Those peo­ple who went to Sil­i­con Val­ley, they were turned around by what Palan­tir had to offer,” said Mortvedt.

The sys­tem was hand­ed over in 2020 but is still not func­tion­al. Palan­tir said that the prob­lems were “not a func­tion of our col­lab­o­ra­tion and, to the best of our knowl­edge, have their root cause else­where.”

The Nor­we­gian police con­firmed that Omnia has cost 93m Nor­we­gian kro­ner, or slight­ly less than €10m.

Palan­tir met Dan­ish offi­cials in Sil­i­con Val­ley two years ear­li­er than their Nor­we­gian coun­ter­parts. The Danes end­ed up buy­ing Gotham for both the police and intel­li­gence ser­vices as part of a counter-ter­ror­ism dri­ve.

Chris­t­ian Svan­berg, who would become the data pro­tec­tion offi­cer for the sys­tem, named POL-INTEL, said he wrote the rel­e­vant leg­is­la­tion enabling POL-INTEL.

The ten­der, which was made pub­lic, called for a sys­tem with cross-cut­ting access to exist­ing police and intel­li­gence data­bas­es, infor­ma­tion exchange with Europol and open-source col­lec­tion of new infor­ma­tion. It also fore­saw the need for algo­rithms to pro­vide pat­tern recog­ni­tion and social media analy­sis.

It was, in oth­er words, a pre­scrip­tion for a pre­dic­tive polic­ing sys­tem, which ven­dors claim can help police pre­dict where crimes will occur (place-based) and who might com­mit them (per­son-based). One of Denmark’s dis­trict police chiefs called it a “quan­tum leap into mod­ern polic­ing”.

Palan­tir said it under­stood from the Dan­ish police that they did not use POL-INTEL for pre­dic­tive polic­ing.

Dan­ish author­i­ties pro­nounce them­selves hap­py with the per­for­mance of POL-INTEL but have so far refused to release an inter­nal eval­u­a­tion or dis­close data to enable any inde­pen­dent assess­ment of the results.

The police have refused to dis­close even redact­ed ver­sions of the inter­nal eval­u­a­tions of POL-INTEL. Despite Dan­ish insis­tence on pri­va­cy safe­guards with POL-INTEL, the only known inter­nal assess­ment of the sys­tem found that police users had been using it to spy on the where­abouts of for­mer Arse­nal foot­baller, Nick­las Bendt­ner. A num­ber of police offi­cers were dis­ci­plined over the mat­ter.

Nor­way and Den­mark were not alone in the enthu­si­asm of their senior police for pre­dic­tive polic­ing, the Ger­many state of Hesse pur­chased a sim­i­lar tool from Palan­tir in a ten­der that the oppo­si­tion in the state par­lia­ment con­sid­ered to be so opaque that a com­mit­tee of inquiry dealt with it.

A Ger­man police offi­cial famil­iar with the devel­op­ment of pre­dic­tive tools at the time says that senior offi­cers had bought into the hype: “What was pro­mot­ed three years ago was the hottest shit ever in polic­ing. What we got wasn’t what was expect­ed. You can’t pre­dict crime.”

The Inte­ri­or Min­istry in Hesse said: “The Hes­s­ian police has had con­sis­tent­ly pos­i­tive expe­ri­ences in its coop­er­a­tion with Palan­tir.”

A bunker in The Hague

Since the EU passed its GDPR leg­is­la­tion in 2018, set­ting a glob­al stan­dard for the pri­va­cy rights of its cit­i­zens, it has talked itself up as a safe haven where dig­i­tal rights are pro­tect­ed as human rights. While GDPR may still be poor­ly under­stood and main­ly asso­ci­at­ed with brows­er requests to accept cook­ies, there is a watch­dog. The Euro­pean data pro­tec­tion super­vi­sor and his staff of 75 face the immense task of ensur­ing that Euro­pean agen­cies and the pri­vate com­pa­nies they con­tract play by the rules. The super­vi­sor him­self is Pol­ish lawyer Woj­ciech Wiewiórows­ki, who led the inspec­tions at Europol pre­vi­ous­ly. Pre­dictably cau­tious in his choice of words, he stops short of call­ing for con­tro­ver­sial com­pa­nies such as Palan­tir to be kept away from sen­si­tive Euro­pean data. But he does coun­sel cau­tion.

“It doesn’t make a dif­fer­ence if sys­tems have been pro­duced in the EU or out­side of it when con­sid­er­ing their com­pli­ance with data pro­tec­tion require­ments. But soft­ware pro­duced by com­pa­nies that might have con­nec­tions with intel­li­gence ser­vices of coun­tries out­side the EU should be of spe­cial inter­est for us.”

It is not always clear who is tak­ing more inter­est in who. Palan­tir has shown it has reach and influ­ence over the shap­ing of knowl­edge around data and pri­va­cy in Europe. Some of the continent’s lead­ing thinkers on big data, arti­fi­cial intel­li­gence and ethics have worked with the com­pa­ny in a paid capac­i­ty. One of them is Nico van Eijk, who held a pro­fes­sor­ship at the Uni­ver­si­ty of Ams­ter­dam. Meet­ing Van Eijk in his cur­rent job is an involved process. These days his office is in a bunker in The Hague in the same build­ing as the Nether­lands’ Coun­cil of State. It is here that he runs the com­mit­tee that over­sees the Dutch intel­li­gence ser­vices.

You can only enter if you leave all dig­i­tal devices at the entrance – no phones, lap­tops, no record­ing devices. Through­out the Covid cri­sis employ­ees could not work from home as their com­mu­ni­ca­tions can­not be trust­ed to an inter­net con­nec­tion. The com­mit­tee has real-time access to all data and inves­ti­ga­tions by the mil­i­tary and gen­er­al intel­li­gence ser­vices of the Nether­lands.

At a meet­ing in Jan­u­ary 2021, Van Eijk declined to dis­cuss a pre­vi­ous role he held on Palantir’s advi­so­ry board but com­mend­ed the com­pa­ny on hav­ing an eth­i­cal board in the first place. Palan­tir said Van Eijk was an advis­er on pri­va­cy and civ­il lib­er­ties and that board mem­bers are “nei­ther asked nor expect­ed to agree with or endorse deci­sions made by Palan­tir” and are “com­pen­sat­ed for their time”.

Cor­po­ra­tions, includ­ing those in tech indus­try, are spon­sor­ing an increas­ing num­ber of aca­d­e­mics with poten­tial impli­ca­tions for the pro­duc­tion of knowl­edge on data and pri­va­cy.

Many of Van Eijk’s col­leagues at the Uni­ver­si­ty of Ams­ter­dam take a dif­fer­ent view of Palan­tir. Ahead of the 2018 Ams­ter­dam Pri­va­cy Con­fer­ence (APC), one of Europe’s pre­mier events on the sub­ject, more than 100 lead­ing schol­ars signed a com­plaint that stat­ed: “The pres­ence of Palan­tir as a spon­sor of this con­fer­ence legit­imis­es the company’s prac­tices and gives it the oppor­tu­ni­ty to posi­tion itself as part of the agen­da … Palantir’s busi­ness mod­el is based on a par­tic­u­lar form of sur­veil­lance cap­i­tal­ism that tar­gets mar­gin­alised com­mu­ni­ties and accel­er­ates the use of dis­crim­i­na­to­ry tech­nolo­gies such as pre­dic­tive polic­ing.”

Palan­tir said it is not a sur­veil­lance com­pa­ny. “We do not pro­vide data col­lec­tion ser­vices, includ­ing tools that enable sur­veil­lance of indi­vid­ual cit­i­zens or con­sumers.”

Infe­ri­or­i­ty com­plex

Europe’s depen­dence on US tech is not a mat­ter of con­cern only for human rights advo­cates and pri­va­cy schol­ars. Some of the biggest busi­ness­es in Ger­many and France have been in talks over the cre­ation of some­thing akin to a safe haven for their own com­mer­cial­ly sen­si­tive data. Those dis­cus­sions revealed that Ger­man car man­u­fac­tur­ers were just as ner­vous as any pri­va­cy cam­paign­er about releas­ing their data to US cloud ser­vices, such as Ama­zon Web Ser­vices.

Mari­et­je Schaake, the direc­tor of Stanford’s Cyber Pol­i­cy Cen­tre, warned that Europe’s “tech infe­ri­or­i­ty com­plex” was lead­ing to bad deci­sions: “We’re build­ing a soft­ware house of cards which is sold as a ser­vice to the pub­lic but can be a lia­bil­i­ty to soci­ety. There’s an asym­me­try of knowl­edge and pow­er and account­abil­i­ty, a ques­tion of what we’re able to know in the pub­lic inter­est. Pri­vate pow­er over pub­lic process­es is grow­ing expo­nen­tial­ly with access to data and tal­ent.”

Palan­tir says that “it suc­cess­ful­ly oper­ates with­in and pro­motes the goals of the GDPR and its under­ly­ing prin­ci­ples”. It insists it is not a data com­pa­ny but rather a soft­ware com­pa­ny that pro­vides data man­age­ment plat­forms. It has for a decade, it says, worked in Europe with com­mer­cial and gov­ern­ment organ­i­sa­tions, “help­ing them suc­cess­ful­ly meet data pro­tec­tion require­ments at scale as man­dat­ed at a Euro­pean and nation­al lev­el”.

The lat­est Euro­pean bid for greater dig­i­tal sov­er­eign­ty is GAIA‑X, wrong­ly billed in some quar­ters as a project to make a Euro-cloud. It is, in fact, an asso­ci­a­tion that will seek to set the rules by which Europe-based com­pa­nies do busi­ness with cloud com­put­ing ser­vices. Just as GDPR means that Euro­peans’ per­son­al data has to be treat­ed dif­fer­ent­ly on Face­book than that of users out­side the EU, GAIA‑X would mean com­mer­cial data is more tight­ly con­trolled on the cloud. Despite its rel­a­tive obscu­ri­ty, GAIA‑X may go on to have pro­found impli­ca­tions for the busi­ness mod­el of US tech com­pa­nies, or hyper­scalers.

It was a sur­prise there­fore when Palan­tir pro­claimed itself, among oth­er com­pa­nies, a “day 1 part­ner” of GAIA‑X three months before any deci­sion had been made. Offi­cials at the asso­ci­a­tion com­plained of “delin­quent part­ners” who had jumped the gun for rea­sons of com­mer­cial advan­tage. Ulti­mate­ly, Palan­tir was allowed to join.

Palan­tir says it did noth­ing that oth­er com­pa­nies involved with GAIA‑X did not do.

The chair­man of GAIA‑X, Hubert Tardieu, for­mer­ly a senior exec­u­tive at French tech firm ATOS, not­ed that the asso­ci­a­tion did not want to get mired in law­suits from “com­pa­nies in Cal­i­for­nia who know a lot about antitrust law.”

Get ready. It’s com­ing. What’s com­ing? We don’t know. And it’s unclear Palan­tir knows. But fol­low­ing reports that Palan­tir just pur­chased $50.7 mil­lion in gold bars and announced that it’s now accept­ing pay­ments in both gold and bit­coin for its soft­ware in antic­i­pa­tion of anoth­er “black swan event”, we have to ask: what is Palan­tir see­ing that they aren’t telling us? What­ev­er it is, it does­n’t appear to bode well for the US. At least not the dol­lar. The move comes rough­ly a year after the com­pa­ny relo­cat­ed from San Fran­cis­co to Den­ver.

This is prob­a­bly a good time to recall that Pres­i­dent Biden’s Direc­tor of Nation­al Intel­li­gence, Avril Haines, was a Palan­tir con­sul­tant from July 5, 2017 to June 23, 2020, when she left to join the Biden cam­paign. It’s a reminder that Palan­tir’s intel­li­gence assess­ments prob­a­bly include plen­ty of infor­ma­tion flows from the numer­ous peo­ple in the intel­li­gence com­mu­ni­ty with ties to the com­pa­ny.

At the same time, it’s worth keep­ing in mind that when a com­pa­ny known for its threat analy­sis capa­bil­i­ties makes big pub­lic pur­chas­es like this, that’s kind of an adver­tise­ment for Palan­tir’s ser­vices. We could be look­ing at some cre­ative mar­ket­ing tac­tics. Either way, for the com­pa­ny that’s effec­tive a pri­va­tized NSA it’s quite a sig­nal to send to the world:

Bloomberg

Palan­tir Buys Gold Bars as Hedge Against ‘Black Swan Event’

* Com­pa­ny spent $50.7 mil­lion on 100-ounce gold bars in August
* Cus­tomers can now pay for soft­ware in gold or Bit­coin

By Lizette Chap­man
August 17, 2021, 12:37 PM CDT
Updat­ed on August 17, 2021, 12:56 PM CDT

Palan­tir Tech­nolo­gies Inc. said it’s prepar­ing for anoth­er “black swan event” by stock­pil­ing gold bars and invit­ing cus­tomers to pay for its data analy­sis soft­ware in gold.

The com­pa­ny spent $50.7 mil­lion this month on gold, part of an unusu­al invest­ment strat­e­gy that also includes star­tups, blank-check com­pa­nies and pos­si­bly Bit­coin. Palan­tir had pre­vi­ous­ly said it would accept Bit­coin as a form of pay­ment before adding pre­cious met­als more recent­ly.

A spokes­woman for Palan­tir said no one has yet paid in either Bit­coin or gold. Accept­ing non­tra­di­tion­al cur­ren­cies “reflects more of a world­view,” Shyam Sankar, the chief oper­at­ing offi­cer, said in an inter­view. “You have to be pre­pared for a future with more black swan events.”

The gold pur­chase was buried in a secu­ri­ties fil­ing last week for its quar­ter­ly finan­cial results and report­ed ear­li­er this week by Barron’s. The accep­tance of gold as a form of pay­ment hasn’t been pre­vi­ous­ly report­ed.

Palantir’s 100-ounce gold bars will be kept in a secure loca­tion in the north­east­ern U.S., accord­ing to the fil­ing. “The com­pa­ny is able to take phys­i­cal pos­ses­sion of the gold bars stored at the facil­i­ty at any time with rea­son­able notice,” Palan­tir wrote.

Palan­tir, co-found­ed by the tech­nol­o­gy bil­lion­aire Peter Thiel and Chief Exec­u­tive Offi­cer Alex Karp, makes soft­ware used by gov­ern­ments and busi­ness­es. It fash­ions itself as a com­pa­ny of free thinkers. Palan­tir relo­cat­ed to Den­ver last year and mocked its peers in Sil­i­con Val­ley on the way out. In the inter­view, Shyam com­pared Palantir’s cul­ture with an “artist colony,” rather than a tech com­pa­ny churn­ing out soft­ware on an assem­bly line.

Gov­ern­ments have strong­ly embraced Palan­tir soft­ware to help them make sense of the coro­n­avirus pan­dem­ic, the cur­rent so-called black swan, a ran­dom and unpre­dictable event.

...

———–

“Palan­tir Buys Gold Bars as Hedge Against ‘Black Swan Event’” by Lizette Chap­man; Bloomberg; 08/17/2021

“The com­pa­ny spent $50.7 mil­lion this month on gold, part of an unusu­al invest­ment strat­e­gy that also includes star­tups, blank-check com­pa­nies and pos­si­bly Bit­coin. Palan­tir had pre­vi­ous­ly said it would accept Bit­coin as a form of pay­ment before adding pre­cious met­als more recent­ly.”

Yeah, it’s cer­tain­ly an unusu­al invest­ment strat­e­gy. And note the expla­na­tion for this unusu­al strat­e­gy, accord­ing to the com­pa­ny’s COO: it’s not that there’s a spe­cif­ic black swan event. It “reflects more a world­view”, where “You have to be pre­pared for a future with more black swan events”:

...
A spokes­woman for Palan­tir said no one has yet paid in either Bit­coin or gold. Accept­ing non­tra­di­tion­al cur­ren­cies “reflects more of a world­view,” Shyam Sankar, the chief oper­at­ing offi­cer, said in an inter­view. “You have to be pre­pared for a future with more black swan events.”
...

And that’s pos­si­ble the most omi­nous answer we could have received. There’s no spe­cif­ic black swan event the com­pa­ny is pro­tect­ing against. Instead, it seems the com­pa­ny has adopt­ed a world­view that assumes a high­er rate of black swan events in the future. A world­view root­ed in a deep­en­ing sense of fore­bod­ing doom.

Although who knows, maybe there is some­thing very spe­cif­ic the com­pa­ny is prepar­ing against. It’s not like they would tell us. Well, oth­er than indi­rect­ly telling us maybe through weird pub­lic invest­ment strate­gies like this.

Just how much data is Ama­zon col­lect­ing on us? That was the ques­tion asked in a new Reuters report when a group of sev­en reporters request from Ama­zon pro­files of all of the infor­ma­tion the com­pa­ny has on them, tak­ing advan­tage of a new fea­ture Ama­zon began mak­ing avail­able to US cus­tomers in ear­ly 2020 after fail­ing to defeat a 2018 Cal­i­for­nia bal­lot mea­sure requir­ing such dis­clo­sures.

This is far from the first time these kinds of ques­tions have been asked about Ama­zon’s high­ly inva­sive prod­ucts designed for areas like the bed­room. Recall how Ama­zon’s Echo device — which comes with cam­eras and an AI — was cap­tur­ing incred­i­ble amounts of infor­ma­tion that was poten­tial­ly be sold to third par­ties. So as we might expect, the reporters who request­ed their data sum­maries were for a bit of a shock when they got the stun­ning­ly detailed array of infor­ma­tion on the reporters and their fam­i­lies gath­ered from all the dif­fer­ent Ama­zon-sold prod­ucts avail­able, from the Amazon.com web­site to Kin­dle e‑readers, Ring smart door­bells, and Alexa smart­s­peak­ers. Even with their jad­ed expec­ta­tions the reporters were stunned to learn the lev­el of detail col­lect­ed about them. One reporter found Alexa alone was cap­tur­ing rough­ly 70 voice record­ings from their house­hold dai­ly on aver­age for the pri­or three and a half years. And while Ama­zon has long assured cus­tomers that any record­ings it stores only include the ques­tions asked by the user, the reporters found the record­ings often when on for much longer. Ama­zon said it’s work­ing on fix­ing those bugs.

Per­haps the most sur­pris­ing find­ing was cap­tured video of the chil­dren ask­ing Alexa how they could get their par­ents to let them “Play”. Alexa appar­ent­ly retrieved exact­ly this kind of advice from the web­site wik­i­How, advis­ing the chil­dren to refute com­mon par­ent argu­ments such as “too vio­lent,” “too expen­sive”, and “you’re not doing well enough in school.” Ama­zon said it does not own wik­i­How and that Alexa some­times responds to requests with infor­ma­tion from web­sites. So while the infor­ma­tion being cap­tured by Ama­zon’s ubiq­ui­tous prod­ucts is a major part of this sto­ry, there’s also the ques­tion of what kind of infor­ma­tion are their prod­ucts feed­ing the users, in par­tic­u­lar all the kids who have dis­cov­ered that Alexa will act as a child-ally in child-par­ent intra-house­hold strug­gles.

Final­ly, we got an update on Ama­zon’s annu­al reports on how it com­plies with law enforce­ment requests for this kind of data. The update is that Ama­zon is no longer giv­ing that info out. Why the restric­tion? The com­pa­ny explains that it expand­ed its law enforce­ment com­pli­ance report to be a glob­al report now and there­fore it decide to stream­line the data. Yep. A non­sense non-answer. Which is the kind of answer that sug­gests gov­ern­ments are prob­a­bly hav­ing a field day, with shades of the NSO Group sto­ry here. So while the main sto­ry here is about the col­lec­tion of all of this pri­vate data by Ama­zon, we can’t for­get that there’s noth­ing stop­ping Ama­zon from shar­ing that data, espe­cial­ly with the gov­ern­ments that it needs per­mis­sion from to con­tin­ue oper­at­ing:

Reuters
Tech­nol­o­gy

A look at the inti­mate details Ama­zon knows about us

By Chris Kirkham and Jef­frey Dastin
Novem­ber 19, 2021 5:35 PM UTC Updat­ed

Nov 19 (Reuters) — As a Vir­ginia law­mak­er, Ibra­heem Sami­rah has stud­ied inter­net pri­va­cy issues and debat­ed how to reg­u­late tech firms’ col­lec­tion of per­son­al data. Still, he was stunned to learn the full details of the infor­ma­tion Amazon.com Inc (AMZN.O) has col­lect­ed on him.

The e‑commerce giant had more than 1,000 con­tacts from his phone. It had records of exact­ly which part of the Quran that Sami­rah, who was raised as a Mus­lim, had lis­tened to on Dec. 17 of last year. The com­pa­ny knew every search he had made on its plat­form, includ­ing one for books on “pro­gres­sive com­mu­ni­ty orga­niz­ing” and oth­er sen­si­tive health-relat­ed inquiries he thought were pri­vate.

“Are they sell­ing prod­ucts, or are they spy­ing on every­day peo­ple?” asked Sami­rah, a Demo­c­ra­t­ic mem­ber of the Vir­ginia House of Del­e­gates.

Sami­rah was among the few Vir­ginia leg­is­la­tors who opposed an indus­try-friend­ly, Ama­zon-draft­ed state pri­va­cy bill that passed ear­li­er this year. At Reuters’ request, Sami­rah asked Ama­zon to dis­close the data it col­lect­ed on him as a con­sumer.

The com­pa­ny gath­ers a vast array of infor­ma­tion on its U.S. cus­tomers, and it start­ed mak­ing that data avail­able to all upon request ear­ly last year, after try­ing and fail­ing to defeat a 2018 Cal­i­for­nia mea­sure requir­ing such dis­clo­sures. (U.S. Ama­zon cus­tomers can obtain their data by fill­ing out a form on Amazon.com.)

Sev­en Reuters reporters also obtained their Ama­zon files. The data reveals the company’s abil­i­ty to amass strik­ing­ly inti­mate por­traits of indi­vid­ual con­sumers.

Ama­zon col­lects data on con­sumers through its Alexa voice assis­tant, its e‑commerce mar­ket­place, Kin­dle e‑readers, Audi­ble audio­books, its video and music plat­forms, home-secu­ri­ty cam­eras and fit­ness track­ers. Alexa-enabled devices make record­ings inside people’s homes, and Ring secu­ri­ty cam­eras cap­ture every vis­i­tor.

Such infor­ma­tion can reveal a person’s height, weight and health; their eth­nic­i­ty (via clues con­tained in voice data) and polit­i­cal lean­ings; their read­ing and buy­ing habits; their where­abouts on any giv­en day, and some­times whom they have met.

One reporter’s dossier revealed that Ama­zon had col­lect­ed more than 90,000 Alexa record­ings of fam­i­ly mem­bers between Decem­ber 2017 and June 2021 – aver­ag­ing about 70 dai­ly. The record­ings includ­ed details such as the names of the reporter’s young chil­dren and their favorite songs.

Ama­zon cap­tured the chil­dren ask­ing how they could con­vince their par­ents to let them “play,” and get­ting detailed instruc­tions from Alexa on how to con­vince their par­ents to buy them video games. Be ful­ly pre­pared, Alexa advised the kids, to refute com­mon par­ent argu­ments such as “too vio­lent,” “too expen­sive” and “you’re not doing well enough in school.” The infor­ma­tion came from a third-par­ty pro­gram used by Alexa called “wik­i­How” that pro­vides how-to advice from more than 180,000 arti­cles, accord­ing to Amazon’s web­site.

Ama­zon said it does not own wik­i­How, but that Alexa some­times responds to requests with infor­ma­tion from web­sites.

Some record­ings involved con­ver­sa­tions between fam­i­ly mem­bers using Alexa devices to com­mu­ni­cate across dif­fer­ent parts of the house. Sev­er­al record­ings cap­tured chil­dren apol­o­giz­ing to their par­ents after being dis­ci­plined. Oth­ers picked up the chil­dren, ages 7, 9 and 12, ask­ing Alexa ques­tions about terms like “pan­sex­u­al.”

In one record­ing, a child asks: “Alexa, what is a vagi­na?” In anoth­er: “Alexa, what does bondage mean?”

The reporter did not real­ize Ama­zon was stor­ing the record­ings before the com­pa­ny dis­closed the data it tracked on the fam­i­ly.

Ama­zon says its Alexa prod­ucts are designed to record as lit­tle as pos­si­ble, start­ing with the trig­ger word, “Alexa,” and stop­ping when the user’s com­mand ends. The record­ings of the reporter’s fam­i­ly, how­ev­er, some­times cap­tured longer con­ver­sa­tions.

In a state­ment, Ama­zon said it has sci­en­tists and engi­neers work­ing to improve the tech­nol­o­gy and avoid false trig­gers that prompt record­ing. The com­pa­ny said it alerts cus­tomers that record­ings are stored when they set up Alexa accounts.

Ama­zon said it col­lects per­son­al data to improve prod­ucts and ser­vices and cus­tomize them to indi­vid­u­als. Asked about the records of Sami­rah lis­ten­ing to the Quran on Amazon’s audio­books ser­vice, Ama­zon said such data allows cus­tomers to pick up where they left off from a pri­or ses­sion.

The only way for cus­tomers to delete much of this per­son­al data is to close their account, Ama­zon said. The com­pa­ny said it retains some infor­ma­tion, such as pur­chase his­to­ry, after account clo­sure to com­ply with legal oblig­a­tions.

Ama­zon said it allows cus­tomers to adjust their set­tings on voice assis­tants and oth­er ser­vices to lim­it the amount of data col­lect­ed. Alexa users, for instance, can stop Ama­zon from sav­ing their record­ings or have them auto­mat­i­cal­ly delet­ed peri­od­i­cal­ly. And they can dis­con­nect their con­tacts or cal­en­dars from their smart-speak­er devices if they don’t want to use Alexa’s call­ing or sched­ul­ing func­tions.

A cus­tomer can opt out of hav­ing their Alexa record­ings exam­ined, but they must nav­i­gate a series of menus and two warn­ings that say: “If you turn this off, voice recog­ni­tion and new fea­tures may not work well for you.” Asked about the warn­ings, Ama­zon said con­sumers who lim­it data col­lec­tion may not be able to per­son­al­ize some fea­tures, such as music play­back.

Sami­rah, 30, got an Ama­zon Alexa-enabled smart speak­er dur­ing last year’s hol­i­day sea­son. He said he only used it for three days before return­ing it after real­iz­ing it was col­lect­ing record­ings. “It real­ly sketched me out,” he said.

The device had already gath­ered all of his phone con­tacts, part of a fea­ture that allows users to make calls through the device. Ama­zon said Alexa users must give per­mis­sion for the com­pa­ny to access phone con­tacts. Cus­tomers must dis­able access to phone con­tacts, not just delete the Alexa app, in order to delete the records from their Ama­zon account.

Sami­rah said he was also unnerved that Ama­zon had detailed records of his audio­book and Kin­dle read­ing ses­sions. Find­ing infor­ma­tion about his lis­ten­ing to the Quran dis­closed in his Ama­zon file, he said, made Sami­rah think about the his­to­ry of U.S. police and intel­li­gence agen­cies sur­veilling Mus­lims for sus­pect­ed ter­ror­ist links after the attacks of Sept. 11, 2001.

“Why do they need to know that?” he asked. Samirah’s term ends in Jan­u­ary, after he lost a bid for re-elec­tion ear­li­er this year.

At times, law-enforce­ment agen­cies seek data on cus­tomers from tech­nol­o­gy com­pa­nies. Ama­zon dis­clos­es that it com­plies with search war­rants and oth­er law­ful court orders seek­ing data the com­pa­ny keeps on an account, while object­ing to “over­broad or oth­er­wise inap­pro­pri­ate requests.”

Ama­zon data for the three years end­ing in June 2020, the lat­est avail­able, show the com­pa­ny com­plied at least par­tial­ly with 75% of sub­poe­nas, search war­rants and oth­er court orders seek­ing data on U.S. cus­tomers. The com­pa­ny ful­ly com­plied with 38% of those requests.

Ama­zon stopped dis­clos­ing how often it com­plies with such requests last year. Asked why, Ama­zon said it expand­ed the scope of the U.S. report to make it glob­al, and “stream­lined” the infor­ma­tion from each coun­try on law enforce­ment inquiries.

The com­pa­ny said it is oblig­at­ed to com­ply with “valid and bind­ing orders,” but that its goal is to release “the min­i­mum” required by law.

Amazon’s 3,500-word pri­va­cy pol­i­cy, which links to more than 20 oth­er pages relat­ed to pri­va­cy and user set­tings, gives the com­pa­ny wide lat­i­tude to col­lect data. Ama­zon said the pol­i­cy describes its col­lec­tion, use and shar­ing of data “in a way that is easy for con­sumers to under­stand.”

That infor­ma­tion can get quite per­son­al. Amazon’s Kin­dle e‑readers, for instance, pre­cise­ly track a user’s read­ing habits, anoth­er reporter’s Ama­zon data file showed. The dis­clo­sure includ­ed records of more than 3,700 read­ing ses­sions since 2017, includ­ing time­stamped logs – to the mil­lisec­ond – of books read. Ama­zon also tracks words high­light­ed or looked up, pages turned and pro­mo­tions seen.

It showed, for instance, that a fam­i­ly mem­ber read “The Mitchell Sis­ters: A Com­plete Romance Series” on Aug. 8, 2020, from 4:52 p.m. until 7:36 p.m., flip­ping 428 pages.

...

————

“A look at the inti­mate details Ama­zon knows about us” by Chris Kirkham and Jef­frey Dastin; Reuters; 11/19/2021

“One reporter’s dossier revealed that Ama­zon had col­lect­ed more than 90,000 Alexa record­ings of fam­i­ly mem­bers between Decem­ber 2017 and June 2021 – aver­ag­ing about 70 dai­ly. The record­ings includ­ed details such as the names of the reporter’s young chil­dren and their favorite songs.”

70 record­ings dai­ly. That’s what Alexa alone was cap­tur­ing in one reporter’s house­hold. Infor­ma­tion from the whole spec­trum of Ama­zon prod­ucts are col­lat­ed into a sin­gle cus­tomer record, gath­er­ing infor­ma­tion every­thing from the Amazon.com web­site search­es and pur­chas­es (some­thing we should expect to be tracked) down to the words high­light­ed in your Kin­dle e‑reader (some­thing one would prob­a­bly not be expect­ed to assume was hap­pen­ing). But what is arguably the most scan­dalous aspect of this sit­u­a­tion is that the reporter was just learn­ing about these dai­ly cap­tures for the first time after it had been going on for near­ly four years. Yes, Ama­zon tech­ni­cal­ly dis­clos­es all of this infor­ma­tion cap­ture, but that’s all part of the scan­dal. Accord­ing to the rules of com­merce, you can appar­ent­ly col­lect what­ev­er infor­ma­tion you want on cus­tomers as long as you tuck away a dis­clo­sure of that data cap­ture some­where in the mas­sive pri­va­cy pol­i­cy:

...
Ama­zon col­lects data on con­sumers through its Alexa voice assis­tant, its e‑commerce mar­ket­place, Kin­dle e‑readers, Audi­ble audio­books, its video and music plat­forms, home-secu­ri­ty cam­eras and fit­ness track­ers. Alexa-enabled devices make record­ings inside people’s homes, and Ring secu­ri­ty cam­eras cap­ture every vis­i­tor.

Such infor­ma­tion can reveal a person’s height, weight and health; their eth­nic­i­ty (via clues con­tained in voice data) and polit­i­cal lean­ings; their read­ing and buy­ing habits; their where­abouts on any giv­en day, and some­times whom they have met.

...

The reporter did not real­ize Ama­zon was stor­ing the record­ings before the com­pa­ny dis­closed the data it tracked on the fam­i­ly.

Ama­zon says its Alexa prod­ucts are designed to record as lit­tle as pos­si­ble, start­ing with the trig­ger word, “Alexa,” and stop­ping when the user’s com­mand ends. The record­ings of the reporter’s fam­i­ly, how­ev­er, some­times cap­tured longer con­ver­sa­tions.

In a state­ment, Ama­zon said it has sci­en­tists and engi­neers work­ing to improve the tech­nol­o­gy and avoid false trig­gers that prompt record­ing. The com­pa­ny said it alerts cus­tomers that record­ings are stored when they set up Alexa accounts.

...

Amazon’s 3,500-word pri­va­cy pol­i­cy, which links to more than 20 oth­er pages relat­ed to pri­va­cy and user set­tings, gives the com­pa­ny wide lat­i­tude to col­lect data. Ama­zon said the pol­i­cy describes its col­lec­tion, use and shar­ing of data “in a way that is easy for con­sumers to under­stand.”

That infor­ma­tion can get quite per­son­al. Amazon’s Kin­dle e‑readers, for instance, pre­cise­ly track a user’s read­ing habits, anoth­er reporter’s Ama­zon data file showed. The dis­clo­sure includ­ed records of more than 3,700 read­ing ses­sions since 2017, includ­ing time­stamped logs – to the mil­lisec­ond – of books read. Ama­zon also tracks words high­light­ed or looked up, pages turned and pro­mo­tions seen.

It showed, for instance, that a fam­i­ly mem­ber read “The Mitchell Sis­ters: A Com­plete Romance Series” on Aug. 8, 2020, from 4:52 p.m. until 7:36 p.m., flip­ping 428 pages.
...

And then there’s this remark­able anec­dote about the kinds of ques­tions chil­dren are pos­ing to Alexa: kids were lit­er­al­ly get­ting advice on how to argue with their par­ents from a web­site that Alexa was access­ing. And then, of course, this was all record­ed. It’s an iron­ic indi­ca­tion of the scale of the poten­tial scan­dal here: the com­pa­ny is lit­er­al­ly record­ing so much data it’s cap­tur­ing the data on its oth­er abus­es:

...
Ama­zon cap­tured the chil­dren ask­ing how they could con­vince their par­ents to let them “play,” and get­ting detailed instruc­tions from Alexa on how to con­vince their par­ents to buy them video games. Be ful­ly pre­pared, Alexa advised the kids, to refute com­mon par­ent argu­ments such as “too vio­lent,” “too expen­sive” and “you’re not doing well enough in school.” The infor­ma­tion came from a third-par­ty pro­gram used by Alexa called “wik­i­How” that pro­vides how-to advice from more than 180,000 arti­cles, accord­ing to Amazon’s web­site.

Ama­zon said it does not own wik­i­How, but that Alexa some­times responds to requests with infor­ma­tion from web­sites.

Some record­ings involved con­ver­sa­tions between fam­i­ly mem­bers using Alexa devices to com­mu­ni­cate across dif­fer­ent parts of the house. Sev­er­al record­ings cap­tured chil­dren apol­o­giz­ing to their par­ents after being dis­ci­plined. Oth­ers picked up the chil­dren, ages 7, 9 and 12, ask­ing Alexa ques­tions about terms like “pan­sex­u­al.”

In one record­ing, a child asks: “Alexa, what is a vagi­na?” In anoth­er: “Alexa, what does bondage mean?”
...

Final­ly, note that when it comes to poten­tial abus­es of this cap­tured data, it’s in the trans­fer of that data to gov­ern­ment agen­cies where the dam­age can real­ly explode. This is a glob­al­ly sold prod­uct, after all. It’s not just going to the US nation­al secu­ri­ty state that’s like­ly get­ting access this all of this incred­i­bly pri­vate data. Pret­ty much any gov­ern­ment is going to poten­tial­ly have a right to request access to it under cer­tain cir­cum­stances. What kind of cir­cum­stances? Well, that pre­sum­ably depends in part on the local laws. That’s all part of why Ama­zon’s deci­sion last year to stop dis­clos­ing how often it com­plies with US law enforce­ment requests is poten­tial­ly so alarm­ing. Ama­zon’s expla­na­tion for end­ing the report was that it expand­ed the report to include shar­ing com­pli­ance glob­al­ly and there­fore stream­lined the avail­able infor­ma­tion. It’s not exact­ly a com­pelling expla­na­tion. So how much of this data is being shared with dif­fer­ent gov­ern­ments? We don’t know, oth­er than we can be pret­ty sure it’s enough to embar­rass Ama­zon into ‘stream­lin­ing’ its reports and lim­it­ing that info:

...
Ama­zon data for the three years end­ing in June 2020, the lat­est avail­able, show the com­pa­ny com­plied at least par­tial­ly with 75% of sub­poe­nas, search war­rants and oth­er court orders seek­ing data on U.S. cus­tomers. The com­pa­ny ful­ly com­plied with 38% of those requests.

Ama­zon stopped dis­clos­ing how often it com­plies with such requests last year. Asked why, Ama­zon said it expand­ed the scope of the U.S. report to make it glob­al, and “stream­lined” the infor­ma­tion from each coun­try on law enforce­ment inquiries.

The com­pa­ny said it is oblig­at­ed to com­ply with “valid and bind­ing orders,” but that its goal is to release “the min­i­mum” required by law.
...

Did Ama­zon “stream­line” the infor­ma­tion on how often it com­plies with US law enforce­ment requests right out of its reports out of a sense of cus­tomer con­ve­nience? That’s the absurd sto­ry the com­pa­ny is telling us. It’s not a great sign.

So the over­all Reuters update on the col­lec­tion of per­son­al infor­ma­tion by Ama­zon appears to be that it is indeed worse than pre­vi­ous­ly rec­og­nized. Which is about as bad as we should have expect­ed.

Right-wing out­rage over ‘Big Tech cen­sor­ship’ of con­ser­v­a­tive voic­es has long been a faith argu­ment made in the spir­it of ‘work­ing the refs’ and gaslight­ing. It’s no secret that the social media giants have been repeat­ed­ly caught giv­ing spe­cial treat­ment to right-wing voic­es on their plat­forms and mak­ing spe­cial excep­tions to excuse and facil­i­tate far right dis­in­for­ma­tion. Dis­in­for­ma­tion that syn­er­gizes with Big Tech’s algo­rithms that pri­or­i­ty ‘engage­ment’, in par­tic­u­lar the anger and fear-dri­ven engage­ment the far right spe­cial­izes in.

So it’s worth point­ing out that when the GOP has been wag­ing its ‘war on Big Tech’ in recent years — end­less­ly rail­ing against alleged mass cen­sor­ship by treat­ing each indi­vid­ual instance of a con­ser­v­a­tive user’s con­tent being pulled for vio­lat­ing the rules as an exam­ple of polit­i­cal dis­crim­i­na­tion — this isn’t just a cyn­i­cal strat­e­gy designed to give social media plat­forms the ‘space’ to give right-wing users more lenient treat­ment than they were oth­er­wise be receiv­ing. It’s also a strat­e­gy that advo­cates for the unchecked exploita­tion of those prof­it-max­i­miz­ing algo­rithms by the plat­forms them­selves. In oth­er words, if Big Tech ever tru­ly did com­plete­ly cave to these far right demands, and allowed the plat­forms’ algo­rithms to be com­plete­ly unchecked ampli­fiers of ‘engag­ing’ far right con­tent as they have been in the past, that does­n’t just help the far right. It’s also a great way for these social media giants to max­i­mize their prof­its.

It’s long been clear that Big Tech and the GOP are play­ing some sort of cyn­i­cal game of polit­i­cal foot­sie with all of these pho­ny ‘Big Tech is cen­sor­ing us’ memes. It’s win-win. The GOP can pre­tend to take a pop­ulist stance on some­thing and Big Tech can pre­tend it’s actu­al­ly doing some­thing to ade­quate­ly address the fact that its plat­forms remain the key tools of fas­cist pol­i­tics glob­al­ly. But giv­en how this con­ser­v­a­tive polit­i­cal cam­paign is lit­er­al­ly fight­ing for Big Tech’s right to oper­ate in an uncheck prof­it-max­i­miz­ing man­ner, we have to ask: just how much secret coor­di­na­tion is there between the GOP and Big Tech in cre­at­ing and orches­trat­ing the GOP’s anti-Big Tech pro­pa­gan­da? Because as this point, you almost could­n’t come up with a more effec­tive lob­by for max­i­miz­ing Big Tech’s prof­its than the army of Repub­li­can offi­cials claim­ing to be very upset with them:

The New York Times
Opin­ion

You Are the Object of a Secret Extrac­tion Oper­a­tion

By Shoshana Zuboff

Dr. Zuboff is a pro­fes­sor emer­i­tus at Har­vard Busi­ness School and the author of “The Age of Sur­veil­lance Cap­i­tal­ism.”
Nov. 12, 2021

Face­book is not just any cor­po­ra­tion. It reached tril­lion-dol­lar sta­tus in a sin­gle decade by apply­ing the log­ic of what I call sur­veil­lance cap­i­tal­ism — an eco­nom­ic sys­tem built on the secret extrac­tion and manip­u­la­tion of human data — to its vision of con­nect­ing the entire world. Face­book and oth­er lead­ing sur­veil­lance cap­i­tal­ist cor­po­ra­tions now con­trol infor­ma­tion flows and com­mu­ni­ca­tion infra­struc­tures across the world.

These infra­struc­tures are crit­i­cal to the pos­si­bil­i­ty of a demo­c­ra­t­ic soci­ety, yet our democ­ra­cies have allowed these com­pa­nies to own, oper­ate and medi­ate our infor­ma­tion spaces uncon­strained by pub­lic law. The result has been a hid­den rev­o­lu­tion in how infor­ma­tion is pro­duced, cir­cu­lat­ed and act­ed upon. A parade of rev­e­la­tions since 2016, ampli­fied by the whis­tle-blow­er Frances Hau­gen’s doc­u­men­ta­tion and per­son­al tes­ti­mo­ny, bears wit­ness to the con­se­quences of this rev­o­lu­tion.

The world’s lib­er­al democ­ra­cies now con­front a tragedy of the “un-com­mons.” Infor­ma­tion spaces that peo­ple assume to be pub­lic are strict­ly ruled by pri­vate com­mer­cial inter­ests for max­i­mum prof­it. The inter­net as a self-reg­u­lat­ing mar­ket has been revealed as a failed exper­i­ment. Sur­veil­lance cap­i­tal­ism leaves a trail of social wreck­age in its wake: the whole­sale destruc­tion of pri­va­cy, the inten­si­fi­ca­tion of social inequal­i­ty, the poi­son­ing of social dis­course with defac­tu­al­ized infor­ma­tion, the demo­li­tion of social norms and the weak­en­ing of demo­c­ra­t­ic insti­tu­tions.

These social harms are not ran­dom. They are tight­ly cou­pled effects of evolv­ing eco­nom­ic oper­a­tions. Each harm paves the way for the next and is depen­dent on what went before.

There is no way to escape the machine sys­tems that sur­veil us, whether we are shop­ping, dri­ving or walk­ing in the park. All roads to eco­nom­ic and social par­tic­i­pa­tion now lead through sur­veil­lance capitalism’s prof­it-max­i­miz­ing insti­tu­tion­al ter­rain, a con­di­tion that has inten­si­fied dur­ing near­ly two years of glob­al plague.

Will Facebook’s dig­i­tal vio­lence final­ly trig­ger our com­mit­ment to take back the “un-com­mons”? Will we con­front the fun­da­men­tal but long ignored ques­tions of an infor­ma­tion civ­i­liza­tion: How should we orga­nize and gov­ern the infor­ma­tion and com­mu­ni­ca­tion spaces of the dig­i­tal cen­tu­ry in ways that sus­tain and advance demo­c­ra­t­ic val­ues and prin­ci­ples?

Search and Seizure

Face­book as we now know it was fash­ioned from Google’s rib. Mark Zuckerberg’s start-up did not invent sur­veil­lance cap­i­tal­ism. Google did that. In 2000, when only 25 per­cent of the world’s infor­ma­tion was stored dig­i­tal­ly, Google was a tiny start-up with a great search prod­uct but lit­tle rev­enue.

By 2001, in the teeth of the dot-com bust, Google’s lead­ers found their break­through in a series of inven­tions that would trans­form adver­tis­ing. Their team learned how to com­bine mas­sive data flows of per­son­al infor­ma­tion with advanced com­pu­ta­tion­al analy­ses to pre­dict where an ad should be placed for max­i­mum “click through.” Pre­dic­tions were com­put­ed ini­tial­ly by ana­lyz­ing data trails that users unknow­ing­ly left behind in the company’s servers as they searched and browsed Google’s pages. Google’s sci­en­tists learned how to extract pre­dic­tive meta­da­ta from this “data exhaust” and use it to ana­lyze like­ly pat­terns of future behav­ior.

Pre­dic­tion was the first imper­a­tive that deter­mined the sec­ond imper­a­tive: extrac­tion. Lucra­tive pre­dic­tions required flows of human data at unimag­in­able scale. Users did not sus­pect that their data was secret­ly hunt­ed and cap­tured from every cor­ner of the inter­net and, lat­er, from apps, smart­phones, devices, cam­eras and sen­sors. User igno­rance was under­stood as cru­cial to suc­cess. Each new prod­uct was a means to more “engage­ment,” a euphemism used to con­ceal illic­it extrac­tion oper­a­tions.

When asked “What is Google?” the co-founder Lar­ry Page laid it out in 2001, accord­ing to a detailed account by Dou­glas Edwards, Google’s first brand man­ag­er, in his book “I’m Feel­ing Lucky”: “Stor­age is cheap. Cam­eras are cheap. Peo­ple will gen­er­ate enor­mous amounts of data,” Mr. Page said. “Every­thing you’ve ever heard or seen or expe­ri­enced will become search­able. Your whole life will be search­able.”

Instead of sell­ing search to users, Google sur­vived by turn­ing its search engine into a sophis­ti­cat­ed sur­veil­lance medi­um for seiz­ing human data. Com­pa­ny exec­u­tives worked to keep these eco­nom­ic oper­a­tions secret, hid­den from users, law­mak­ers, and com­peti­tors. Mr. Page opposed any­thing that might “stir the pri­va­cy pot and endan­ger our abil­i­ty to gath­er data,” Mr. Edwards wrote.

Mas­sive-scale extrac­tion oper­a­tions were the key­stone to the new eco­nom­ic edi­fice and super­seded oth­er con­sid­er­a­tions, begin­ning with the qual­i­ty of infor­ma­tion, because in the log­ic of sur­veil­lance cap­i­tal­ism, infor­ma­tion integri­ty is not cor­re­lat­ed with rev­enue.

This is the eco­nom­ic con­text in which dis­in­for­ma­tion wins. As recent­ly as 2017, Eric Schmidt, the exec­u­tive chair­man of Google’s par­ent com­pa­ny, Alpha­bet, acknowl­edged the role of Google’s algo­rith­mic rank­ing oper­a­tions in spread­ing cor­rupt infor­ma­tion. “There is a line that we can’t real­ly get across,” he said. “It is very dif­fi­cult for us to under­stand truth.” A com­pa­ny with a mis­sion to orga­nize and make acces­si­ble all the world’s infor­ma­tion using the most sophis­ti­cat­ed machine sys­tems can­not dis­cern cor­rupt infor­ma­tion.

Face­book, the First Fol­low­er

Mr. Zucker­berg began his entre­pre­neur­ial career in 2003 while a stu­dent at Har­vard. His web­site, Face­mash, invit­ed vis­i­tors to rate oth­er stu­dents’ attrac­tive­ness. It quick­ly drew out­rage from his peers and was shut­tered. Then came The­Face­book in 2004 and Face­book in 2005, when Zucker­berg acquired his first pro­fes­sion­al investors.

Facebook’s user num­bers quick­ly grew; its rev­enues did not. Like Google a few years ear­li­er, Mr. Zucker­berg could not turn pop­u­lar­i­ty into prof­it. Instead, he careened from blun­der to blun­der. His crude vio­la­tions of users’ pri­va­cy expec­ta­tions pro­voked intense pub­lic back­lash, peti­tions and class-action suits. Mr. Zucker­berg seemed to under­stand that the answer to his prob­lems involved human data extrac­tion with­out con­sent for the sake of adver­tis­ers’ advan­tage, but the com­plex­i­ties of the new log­ic elud­ed him.

...

In March 2008, Mr. Zucker­berg hired Google’s head of glob­al online adver­tis­ing, Sheryl Sand­berg, as his sec­ond in com­mand. Ms. Sand­berg had joined Google in 2001 and was a key play­er in the sur­veil­lance cap­i­tal­ism rev­o­lu­tion. She led the build-out of Google’s adver­tis­ing engine, AdWords, and its AdSense pro­gram, which togeth­er account­ed for most of the company’s $16.6 bil­lion in rev­enue in 2007.

A Google mul­ti­mil­lion­aire by the time she met Mr. Zucker­berg, Ms. Sand­berg had a can­ny appre­ci­a­tion of Facebook’s immense oppor­tu­ni­ties for extrac­tion of rich pre­dic­tive data. “We have bet­ter infor­ma­tion than any­one else. We know gen­der, age, loca­tion, and it’s real data as opposed to the stuff oth­er peo­ple infer,” Ms. Sand­berg explained, accord­ing to David Kirk­patrick in “The Face­book Effect.”

The com­pa­ny had “bet­ter data” and “real data” because it had a front-row seat to what Mr. Page had called “your whole life.”

Face­book paved the way for sur­veil­lance eco­nom­ics with new pri­va­cy poli­cies in late 2009. The Elec­tron­ic Fron­tier Foun­da­tion warned that new “Every­one” set­tings elim­i­nat­ed options to restrict the vis­i­bil­i­ty of per­son­al data, instead treat­ing it as pub­licly avail­able infor­ma­tion.

TechCrunch sum­ma­rized the corporation’s strat­e­gy: “Face­book is forc­ing users to choose their new pri­va­cy options to pro­mote the ‘Every­one’ update, and to clear itself of any poten­tial wrong­do­ing going for­ward. If there is sig­nif­i­cant back­lash against the social net­work, it can claim that users will­ing­ly made the choice to share their infor­ma­tion with every­one.”

Weeks lat­er, Mr. Zucker­berg defend­ed these moves to a TechCrunch inter­view­er. “A lot of com­pa­nies would be trapped by the con­ven­tions and their lega­cies,” he boast­ed. “We decid­ed that these would be the social norms now, and we just went for it.”

Mr. Zucker­berg “just went for it” because there were no laws to stop him from join­ing Google in the whole­sale destruc­tion of pri­va­cy. If law­mak­ers want­ed to sanc­tion him as a ruth­less prof­it-max­i­miz­er will­ing to use his social net­work against soci­ety, then 2009 to 2010 would have been a good oppor­tu­ni­ty.

A Sweep­ing Eco­nom­ic Order

Face­book was the first fol­low­er, but not the last. Google, Face­book, Ama­zon, Microsoft and Apple are pri­vate sur­veil­lance empires, each with dis­tinct busi­ness mod­els. Google and Face­book are data com­pa­nies and sur­veil­lance-cap­i­tal­ist pure plays. The oth­ers have var­ied lines of busi­ness that may include data, ser­vices, soft­ware and phys­i­cal prod­ucts. In 2021 these five U.S. tech giants rep­re­sent five of the six largest pub­licly trad­ed com­pa­nies by mar­ket cap­i­tal­iza­tion in the world.

As we move into the third decade of the 21st cen­tu­ry, sur­veil­lance cap­i­tal­ism is the dom­i­nant eco­nom­ic insti­tu­tion of our time. In the absence of coun­ter­vail­ing law, this sys­tem suc­cess­ful­ly medi­ates near­ly every aspect of human engage­ment with dig­i­tal infor­ma­tion. The promise of the sur­veil­lance div­i­dend now draws sur­veil­lance eco­nom­ics into the “nor­mal” econ­o­my, from insur­ance, retail, bank­ing and finance to agri­cul­ture, auto­mo­biles, edu­ca­tion, health care and more. Today all apps and soft­ware, no mat­ter how benign they appear, are designed to max­i­mize data col­lec­tion.

His­tor­i­cal­ly, great con­cen­tra­tions of cor­po­rate pow­er were asso­ci­at­ed with eco­nom­ic harms. But when human data are the raw mate­r­i­al and pre­dic­tions of human behav­ior are the prod­uct, then the harms are social rather than eco­nom­ic. The dif­fi­cul­ty is that these nov­el harms are typ­i­cal­ly under­stood as sep­a­rate, even unre­lat­ed, prob­lems, which makes them impos­si­ble to solve. Instead, each new stage of harm cre­ates the con­di­tions for the next stage.

All of it begins with extrac­tion. An eco­nom­ic order found­ed on the secret mas­sive-scale extrac­tion of human data assumes the destruc­tion of pri­va­cy as a non­nego­tiable con­di­tion of its busi­ness oper­a­tions. With pri­va­cy out of the way, ill-got­ten human data are con­cen­trat­ed with­in pri­vate cor­po­ra­tions, where they are claimed as cor­po­rate assets to be deployed at will.

The social effect is a new form of inequal­i­ty, reflect­ed in the colos­sal asym­me­try between what these com­pa­nies know about us and what we know about them. The sheer size of this knowl­edge gap is con­veyed in a leaked 2018 Face­book doc­u­ment, which described its arti­fi­cial intel­li­gence hub, ingest­ing tril­lions of behav­ioral data points every day and pro­duc­ing six mil­lion behav­ioral pre­dic­tions each sec­ond.

Next, these human data are weaponized as tar­get­ing algo­rithms, engi­neered to max­i­mize extrac­tion and aimed back at their unsus­pect­ing human sources to increase engage­ment. Tar­get­ing mech­a­nisms change real life, some­times with grave con­se­quences. For exam­ple, the Face­book Files depict Mr. Zucker­berg using his algo­rithms to rein­force or dis­rupt the behav­ior of bil­lions of peo­ple. Anger is reward­ed or ignored. News sto­ries become more trust­wor­thy or unhinged. Pub­lish­ers pros­per or with­er. Polit­i­cal dis­course turns ugli­er or more mod­er­ate. Peo­ple live or die.

Occa­sion­al­ly the fog clears to reveal the ulti­mate harm: the grow­ing pow­er of tech giants will­ing to use their con­trol over crit­i­cal infor­ma­tion infra­struc­ture to com­pete with demo­c­ra­t­i­cal­ly elect­ed law­mak­ers for soci­etal dom­i­nance. Ear­ly in the pan­dem­ic, for exam­ple, Apple and Google refused to adapt their oper­at­ing sys­tems to host con­tact-trac­ing apps devel­oped by pub­lic health author­i­ties and sup­port­ed by elect­ed offi­cials. In Feb­ru­ary, Face­book shut down many of its pages in Aus­tralia as a sig­nal of refusal to nego­ti­ate with the Aus­tralian Par­lia­ment over fees for news con­tent.

That’s why, when it comes to the tri­umph of sur­veil­lance capitalism’s rev­o­lu­tion, it is the law­mak­ers of every lib­er­al democ­ra­cy, espe­cial­ly in the Unit­ed States, who bear the great­est bur­den of respon­si­bil­i­ty. They allowed pri­vate cap­i­tal to rule our infor­ma­tion spaces dur­ing two decades of spec­tac­u­lar growth, with no laws to stop it.

Fifty years ago the con­ser­v­a­tive econ­o­mist Mil­ton Fried­man exhort­ed Amer­i­can exec­u­tives, “There is one and only one social respon­si­bil­i­ty of busi­ness — to use its resources and engage in activ­i­ties designed to increase its prof­its so long as it stays with­in the rules of the game.” Even this rad­i­cal doc­trine did not reck­on with the pos­si­bil­i­ty of no rules.

Democracy’s Coun­ter­rev­o­lu­tion

Demo­c­ra­t­ic soci­eties rived by eco­nom­ic inequal­i­ty, cli­mate cri­sis, social exclu­sion, racism, pub­lic health emer­gency and weak­ened insti­tu­tions have a long climb toward heal­ing. We can’t fix all our prob­lems at once, but we won’t fix any of them, ever, unless we reclaim the sanc­ti­ty of infor­ma­tion integri­ty and trust­wor­thy com­mu­ni­ca­tions. The abdi­ca­tion of our infor­ma­tion and com­mu­ni­ca­tion spaces to sur­veil­lance cap­i­tal­ism has become the meta-cri­sis of every repub­lic, because it obstructs solu­tions to all oth­er crises.

Nei­ther Google, nor Face­book, nor any oth­er cor­po­rate actor in this new eco­nom­ic order set out to destroy soci­ety, any more than the fos­sil fuel indus­try set out to destroy the earth. But like glob­al warm­ing, the tech giants and their fel­low trav­el­ers have been will­ing to treat their destruc­tive effects on peo­ple and soci­ety as col­lat­er­al dam­age — the unfor­tu­nate but unavoid­able byprod­uct of per­fect­ly legal eco­nom­ic oper­a­tions that have pro­duced some of the wealth­i­est and most pow­er­ful cor­po­ra­tions in the his­to­ry of cap­i­tal­ism.

Where does that leave us? Democ­ra­cy is the only coun­ter­vail­ing insti­tu­tion­al order with the legit­i­mate author­i­ty and pow­er to change our course. If the ide­al of human self-gov­er­nance is to sur­vive the dig­i­tal cen­tu­ry, then all solu­tions point to one solu­tion: a demo­c­ra­t­ic coun­ter­rev­o­lu­tion. But instead of the usu­al laun­dry lists of reme­dies, law­mak­ers need to pro­ceed with a clear grasp of the adver­sary: a sin­gle hier­ar­chy of eco­nom­ic caus­es and their social harms.

We can’t rid our­selves of lat­er-stage social harms unless we out­law their foun­da­tion­al eco­nom­ic caus­es. This means we move beyond the cur­rent focus on down­stream issues such as con­tent mod­er­a­tion and polic­ing ille­gal con­tent. Such “reme­dies” only treat the symp­toms with­out chal­leng­ing the ille­git­i­ma­cy of the human data extrac­tion that funds pri­vate con­trol over society’s infor­ma­tion spaces. Sim­i­lar­ly, struc­tur­al solu­tions like “break­ing up” the tech giants may be valu­able in some cas­es, but they will not affect the under­ly­ing eco­nom­ic oper­a­tions of sur­veil­lance cap­i­tal­ism.

Instead, dis­cus­sions about reg­u­lat­ing big tech should focus on the bedrock of sur­veil­lance eco­nom­ics: the secret extrac­tion of human data from realms of life once called “pri­vate.” Reme­dies that focus on reg­u­lat­ing extrac­tion are con­tent neu­tral. They do not threat­en free­dom of expres­sion. Instead, they lib­er­ate social dis­course and infor­ma­tion flows from the “arti­fi­cial selec­tion” of prof­it-max­i­miz­ing com­mer­cial oper­a­tions that favor infor­ma­tion cor­rup­tion over integri­ty. They restore the sanc­ti­ty of social com­mu­ni­ca­tions and indi­vid­ual expres­sion.

No secret extrac­tion means no ille­git­i­mate con­cen­tra­tions of knowl­edge about peo­ple. No con­cen­tra­tions of knowl­edge means no tar­get­ing algo­rithms. No tar­get­ing means that cor­po­ra­tions can no longer con­trol and curate infor­ma­tion flows and social speech or shape human behav­ior to favor their inter­ests. Reg­u­lat­ing extrac­tion would elim­i­nate the sur­veil­lance div­i­dend and with it the finan­cial incen­tives for sur­veil­lance.

While lib­er­al democ­ra­cies have begun to engage with the chal­lenges of reg­u­lat­ing today’s pri­vate­ly owned infor­ma­tion spaces, the sober truth is that we need law­mak­ers ready to engage in a once-a-cen­tu­ry explo­ration of far more basic ques­tions: How should we struc­ture and gov­ern infor­ma­tion, con­nec­tion and com­mu­ni­ca­tion in a demo­c­ra­t­ic dig­i­tal cen­tu­ry? What new char­ters of rights, leg­isla­tive frame­works and insti­tu­tions are required to ensure that data col­lec­tion and use serve the gen­uine needs of indi­vid­u­als and soci­ety? What mea­sures will pro­tect cit­i­zens from unac­count­able pow­er over infor­ma­tion, whether it is wield­ed by pri­vate com­pa­nies or gov­ern­ments?

Lib­er­al democ­ra­cies should take the lead because they have the pow­er and legit­i­ma­cy to do so. But they should know that their allies and col­lab­o­ra­tors include the peo­ple of every soci­ety strug­gling against a dystopi­an future.

...

———–

“You Are the Object of a Secret Extrac­tion Oper­a­tion” by Shoshana Zuboff; The New York Times; 11/12/2021

“This is the eco­nom­ic con­text in which dis­in­for­ma­tion wins. As recent­ly as 2017, Eric Schmidt, the exec­u­tive chair­man of Google’s par­ent com­pa­ny, Alpha­bet, acknowl­edged the role of Google’s algo­rith­mic rank­ing oper­a­tions in spread­ing cor­rupt infor­ma­tion. “There is a line that we can’t real­ly get across,” he said. “It is very dif­fi­cult for us to under­stand truth.” A com­pa­ny with a mis­sion to orga­nize and make acces­si­ble all the world’s infor­ma­tion using the most sophis­ti­cat­ed machine sys­tems can­not dis­cern cor­rupt infor­ma­tion.”

An eco­nom­ic par­a­digm cen­tered on max­i­miz­ing prof­its by pro­cess­ing ever-increas­ing vol­umes of per­son­al infor­ma­tion for the pur­pose of pre­dict­ing user behav­ior. And yet this par­a­digm can’t actu­al­ly dis­cern truth. A giant infor­ma­tion-pro­cess­ing-and-deliv­er­ing sys­tem that can’t deter­mine whether or not the infor­ma­tion its pro­cess­ing or deliv­er is cor­rupt infor­ma­tion. Cor­rupt or not, it’s the col­lec­tion and deliv­ery of infor­ma­tion that max­i­mizes prof­its. Ped­dling dis­in­for­ma­tion is how these com­pa­nies max­i­mize their prof­its. If prof­it-max­i­miz­ing is the over­ar­ch­ing imper­a­tive dri­ving the actions of these enti­ties, the pro­mo­tion dis­in­for­ma­tion is a nec­es­sary con­se­quence. You can’t dis­en­tan­gle the two:

...
The world’s lib­er­al democ­ra­cies now con­front a tragedy of the “un-com­mons.” Infor­ma­tion spaces that peo­ple assume to be pub­lic are strict­ly ruled by pri­vate com­mer­cial inter­ests for max­i­mum prof­it. The inter­net as a self-reg­u­lat­ing mar­ket has been revealed as a failed exper­i­ment. Sur­veil­lance cap­i­tal­ism leaves a trail of social wreck­age in its wake: the whole­sale destruc­tion of pri­va­cy, the inten­si­fi­ca­tion of social inequal­i­ty, the poi­son­ing of social dis­course with defac­tu­al­ized infor­ma­tion, the demo­li­tion of social norms and the weak­en­ing of demo­c­ra­t­ic insti­tu­tions.

These social harms are not ran­dom. They are tight­ly cou­pled effects of evolv­ing eco­nom­ic oper­a­tions. Each harm paves the way for the next and is depen­dent on what went before.

There is no way to escape the machine sys­tems that sur­veil us, whether we are shop­ping, dri­ving or walk­ing in the park. All roads to eco­nom­ic and social par­tic­i­pa­tion now lead through sur­veil­lance capitalism’s prof­it-max­i­miz­ing insti­tu­tion­al ter­rain, a con­di­tion that has inten­si­fied dur­ing near­ly two years of glob­al plague.

...

Pre­dic­tion was the first imper­a­tive that deter­mined the sec­ond imper­a­tive: extrac­tion. Lucra­tive pre­dic­tions required flows of human data at unimag­in­able scale. Users did not sus­pect that their data was secret­ly hunt­ed and cap­tured from every cor­ner of the inter­net and, lat­er, from apps, smart­phones, devices, cam­eras and sen­sors. User igno­rance was under­stood as cru­cial to suc­cess. Each new prod­uct was a means to more “engage­ment,” a euphemism used to con­ceal illic­it extrac­tion oper­a­tions.

...

Mas­sive-scale extrac­tion oper­a­tions were the key­stone to the new eco­nom­ic edi­fice and super­seded oth­er con­sid­er­a­tions, begin­ning with the qual­i­ty of infor­ma­tion, because in the log­ic of sur­veil­lance cap­i­tal­ism, infor­ma­tion integri­ty is not cor­re­lat­ed with rev­enue.

...

We can’t rid our­selves of lat­er-stage social harms unless we out­law their foun­da­tion­al eco­nom­ic caus­es. This means we move beyond the cur­rent focus on down­stream issues such as con­tent mod­er­a­tion and polic­ing ille­gal con­tent. Such “reme­dies” only treat the symp­toms with­out chal­leng­ing the ille­git­i­ma­cy of the human data extrac­tion that funds pri­vate con­trol over society’s infor­ma­tion spaces. Sim­i­lar­ly, struc­tur­al solu­tions like “break­ing up” the tech giants may be valu­able in some cas­es, but they will not affect the under­ly­ing eco­nom­ic oper­a­tions of sur­veil­lance cap­i­tal­ism.

Instead, dis­cus­sions about reg­u­lat­ing big tech should focus on the bedrock of sur­veil­lance eco­nom­ics: the secret extrac­tion of human data from realms of life once called “pri­vate.” Reme­dies that focus on reg­u­lat­ing extrac­tion are con­tent neu­tral. They do not threat­en free­dom of expres­sion. Instead, they lib­er­ate social dis­course and infor­ma­tion flows from the “arti­fi­cial selec­tion” of prof­it-max­i­miz­ing com­mer­cial oper­a­tions that favor infor­ma­tion cor­rup­tion over integri­ty. They restore the sanc­ti­ty of social com­mu­ni­ca­tions and indi­vid­ual expres­sion.
...

It’s that inex­tri­ca­ble nature of the prof­it-max­i­miz­ing motives of these Sil­i­con Val­ley giants and the imper­a­tive to pro­mote mis­in­for­ma­tion that points us towards what ulti­mate must be part of the solu­tion here: acknowl­edg­ing that democ­ra­cy can’t sur­vive in an envi­ron­ment when dis­in­for­ma­tion is algo­rith­mi­cal­ly pro­mot­ed under the cold direc­tive of prof­it max­i­miza­tion. It real­ly is a choice of which sys­tem will ulti­mate­ly reign supreme. Democ­ra­cy or sur­veil­lance cap­i­tal­ism:

...
Occa­sion­al­ly the fog clears to reveal the ulti­mate harm: the grow­ing pow­er of tech giants will­ing to use their con­trol over crit­i­cal infor­ma­tion infra­struc­ture to com­pete with demo­c­ra­t­i­cal­ly elect­ed law­mak­ers for soci­etal dom­i­nance. Ear­ly in the pan­dem­ic, for exam­ple, Apple and Google refused to adapt their oper­at­ing sys­tems to host con­tact-trac­ing apps devel­oped by pub­lic health author­i­ties and sup­port­ed by elect­ed offi­cials. In Feb­ru­ary, Face­book shut down many of its pages in Aus­tralia as a sig­nal of refusal to nego­ti­ate with the Aus­tralian Par­lia­ment over fees for news con­tent.

That’s why, when it comes to the tri­umph of sur­veil­lance capitalism’s rev­o­lu­tion, it is the law­mak­ers of every lib­er­al democ­ra­cy, espe­cial­ly in the Unit­ed States, who bear the great­est bur­den of respon­si­bil­i­ty. They allowed pri­vate cap­i­tal to rule our infor­ma­tion spaces dur­ing two decades of spec­tac­u­lar growth, with no laws to stop it.

...

Demo­c­ra­t­ic soci­eties rived by eco­nom­ic inequal­i­ty, cli­mate cri­sis, social exclu­sion, racism, pub­lic health emer­gency and weak­ened insti­tu­tions have a long climb toward heal­ing. We can’t fix all our prob­lems at once, but we won’t fix any of them, ever, unless we reclaim the sanc­ti­ty of infor­ma­tion integri­ty and trust­wor­thy com­mu­ni­ca­tions. The abdi­ca­tion of our infor­ma­tion and com­mu­ni­ca­tion spaces to sur­veil­lance cap­i­tal­ism has become the meta-cri­sis of every repub­lic, because it obstructs solu­tions to all oth­er crises.

Nei­ther Google, nor Face­book, nor any oth­er cor­po­rate actor in this new eco­nom­ic order set out to destroy soci­ety, any more than the fos­sil fuel indus­try set out to destroy the earth. But like glob­al warm­ing, the tech giants and their fel­low trav­el­ers have been will­ing to treat their destruc­tive effects on peo­ple and soci­ety as col­lat­er­al dam­age — the unfor­tu­nate but unavoid­able byprod­uct of per­fect­ly legal eco­nom­ic oper­a­tions that have pro­duced some of the wealth­i­est and most pow­er­ful cor­po­ra­tions in the his­to­ry of cap­i­tal­ism.

Where does that leave us? Democ­ra­cy is the only coun­ter­vail­ing insti­tu­tion­al order with the legit­i­mate author­i­ty and pow­er to change our course. If the ide­al of human self-gov­er­nance is to sur­vive the dig­i­tal cen­tu­ry, then all solu­tions point to one solu­tion: a demo­c­ra­t­ic coun­ter­rev­o­lu­tion. But instead of the usu­al laun­dry lists of reme­dies, law­mak­ers need to pro­ceed with a clear grasp of the adver­sary: a sin­gle hier­ar­chy of eco­nom­ic caus­es and their social harms.
...

It’s worth recall­ing at this point the reports of the secret din­ner in the fall of the 2019 between Mark Zucker­berg, Peter Thiel, Jared Kush­n­er and Don­ald Trump at the White House dur­ing one of Zucker­berg’s trips to DC. Zucker­berg and Trump appar­ent­ly came to an agree­ment dur­ing the dinne where Zucker­berg promised that Face­book would take a hands-off approach to the polic­ing of mis­in­for­ma­tion from con­ser­v­a­tive sites. So as we see this far­ci­cal spat between the GOP and Big Tech play out to the syn­er­gis­tic ben­e­fit of both the GOP and Big Tech’s investors, we should prob­a­bly be ask­ing what else was agreed upon at that secret meet­ing and the oth­er secret meet­ings that have undoubt­ed­ly been tak­ing place all along between the Sil­i­con Val­ley giants and pow­er­ful forces on the far right. Was this pho­ny GOP-vs-Big Tech cam­paign active­ly dis­cussed out dur­ing that meet­ing? Because as Shoshana Zuboff observes, this real­ly is a choice between democ­ra­cy and max­i­mum prof­its, and it’s pret­ty clear Big Tech and the GOP both made the same choice a while ago.

The dom­i­na­tion in the social media space of com­pa­nies with deep ties to the US mil­i­tary indus­tri­al com­plex is noth­ing new, as Yasha Levine doc­u­ment­ed in his book Sur­veil­lance Val­ley. So with Elon Musk hav­ing just tak­en per­son­al con­trol of Twit­ter, it’s worth not­ing that Musk isn’t just a lib­er­tar­i­an bil­lion­aire who is clear­ly find­ing joy in trolling the left with his new pow­er over this key social media plat­form. As Levine reminds us below, he’s a US defense con­trac­tor and that role is poised to only grow.

It’s a fun fact that adds con­text to Musk’s hyper-troll­ish tweet a cou­ple of days ago of a car­toon depict­ing the clas­sic far right trope that the polar­iza­tion in US pol­i­tics is exclu­sive­ly due to Democ­rats and lib­er­als lurch­ing to the extreme left, push­ing for­mer lib­er­als like Musk into the con­ser­v­a­tive camp. The car­toon shows three stick fig­ures at three dif­fer­ent time peri­ods: in 2008, it’s “my fel­low lib­er­al” on the left, “me” (Musk) in the cen­ter left, and a con­ser­v­a­tive on the right. A 2012 scene shows the “my fel­low lib­er­al” run­ning quick­ly to the left, mov­ing “me” to the cen­ter. Final­ly, there’s a 2021 scene show­ing the lib­er­al far out to the left shout­ing “Big­ot!”, with “me” now in the cen­ter-right part of the plot and the con­ser­v­a­tive stick­fig­ure exclaim­ing “LOL!”. Musk basi­cal­ly came out as a ‘for­mer lib­er­al’ in the tweet.

And as Greg Sar­gent points out in the fol­low­ing piece, that tweet­ed car­toon was­n’t just an expres­sion of Musk’s pol­i­tics. It was basi­cal­ly a state­ment of intent. An intent to allow Twit­ter to revert back into a Alt Right fan­ta­sy plat­form where ‘any­thing goes’ and far right dis­in­for­ma­tion dom­i­nates.

And this is of course all hap­pen­ing in the midst of the GOP’s deep­en­ing embrace of the pol­i­tics of QAnon and insur­rec­tion. At this point, the GOP’s qua­si-offi­cial stance is that the Demo­c­ra­t­ic Par­ty con­sists of ‘groomers’ try­ing to change the law to make it eas­i­er to prey on chil­dren. How is Musk plan­ning on han­dling the inevitable del­uge of tweets pro­mot­ing insur­rec­tion and call­ing for the death of pedophile Democ­rats?

These are the kinds of ques­tions Musk is going to have to answer at some point and based on his pub­lic com­ments thus far it’s not at all clear that he’s thought it through at all. Or maybe he has thought it through and the plan real­ly is to just allow Twit­ter to revert back into an ‘any­thing goes’ plat­form. We’ll see.

At the same time, there are cer­tain­ly some areas where social media plat­forms real­ly could use a loos­en­ing on their mod­er­a­tion poli­cies, in par­tic­u­lar when it comes to glob­al events involv­ing Rus­sia or Chi­na. Recall how Ukrain­ian Jew­ish activist Eduard Dolin­sky was lit­er­al­ly banned from Face­book for show­ing exam­ples of the kind of anti-Semit­ic graf­fi­ti that has become ram­pant in Ukraine. Also recall how Twit­ter itself locked the offi­cial Twit­ter account of the Chi­nese embassy in the US back in Jan­u­ary 2021 over a tweet defend­ing Bei­jing’s treat­ment of Uyghurs. Per­haps Musk can address this kind of cen­sor­ship being done on behalf of the US nation­al secu­ri­ty state. But that returns us to the fact that Musk is very much a US defense con­trac­tor and that rela­tion­ship with the US nation­al secu­ri­ty state is only get­ting deep­er. Musk real­ly is part of ‘the Deep State’. A ‘Deep State’ with that has decades of work­ing rela­tion­ships with far right ele­ments around the globe. But unlike most ele­ments of the Deep State, he’s got a right-wing fan base that seems to fan­cy Musk some sort of fel­low trav­el­er ‘out­sider’. It’s a fas­ci­nat­ing sit­u­a­tion. A fas­ci­nat­ing sit­u­a­tion that does­n’t bode well.

Ok, first, here’s Sar­gen­t’s piece on Musk’s recent tweet where he basi­cal­ly comes out as a repub­li­can. What is the fall out going to be now that Musk is more or less promis­ing to revert Twit­ter back into an ‘any­thing goes’ dis­in­for­ma­tion machine? we’ll find out...probably dur­ing the next insur­rec­tion fueled by waves of retweet­ed deep fake videos por­tray­ing democ­rats as satan­ic pedophiles:

The Wash­ing­ton Post
Opin­ion

Elon Musk reveals how his Twit­ter may fuel right-wing extrem­ism

By Greg Sar­gent
Colum­nist |
04/29/2022 at 11:44 a.m. EDT

Elon Musk, the most pow­er­ful Twit­ter troll in our soon-to-be inter­plan­e­tary species, has spent the past 24 hours trolling Democ­rats. In tweet after tweet, he has claimed the Demo­c­ra­t­ic Par­ty has lurched into left­ist extrem­ism, leav­ing no option for noble cen­trists — a class that coin­ci­den­tal­ly includes Musk him­self — but to turn toward the GOP.

Musk’s core claim has been bru­tal­ly debunked. As observers quick­ly demon­strat­ed, the data shows the oppo­site: By numer­ous met­rics, it’s the GOP that has grown far more extreme.

But it’s a mis­take to eval­u­ate Musk’s trolling pure­ly as an empir­i­cal mat­ter. It might bet­ter be seen as a state­ment of intent. Musk’s trolling might be reveal­ing in real time how his pend­ing pur­chase of Twit­ter will fuel and enable right-wing extrem­ism.

“I strong­ly sup­port­ed Oba­ma for pres­i­dent, but today’s Demo­c­ra­t­ic Par­ty has been hijacked by extrem­ists,” Musk tweet­ed about Barack Oba­ma. He added: “The far left hates every­one, them­selves includ­ed!” And there’s this:

pic.twitter.com/Q9OjlJhi7f
— Elon Musk (@elonmusk) April 28, 2022

The idea that con­ser­v­a­tives haven’t moved right, as depict­ed here, is false by many ide­o­log­i­cal met­rics. But the rub here is not the empir­i­cal fail­ing. It’s the sig­nal of intent to turn a blind eye to the real­i­ties of right-wing extrem­ism. This graph­ic lit­er­al­ly depicts Musk him­self doing this.

That’s only one tweet, and it’s a troll. But his entire pub­lic argu­ment hints at that under­ly­ing aim, with real-world impli­ca­tions.

Musk has vowed to relax Twitter’s mod­er­a­tion poli­cies. He has attacked those who fear what this means as oppo­nents of “free speech,” while sug­gest­ing that those pre­vi­ous poli­cies had a left-wing bias.

It’s not hard to see what that might add up to: It might mean relax­ing mod­er­a­tion of the sort of tweet­ing on the extreme right that had pre­vi­ous­ly been mod­er­at­ed against, in the name of “free speech” and restor­ing ide­o­log­i­cal “bal­ance.”

To be clear, the point here isn’t that mod­er­a­tion poli­cies shouldn’t be rethought. Sure­ly some mod­er­a­tion of par­tic­u­lar right-wing speech has been exces­sive. If Musk can make the mod­er­a­tion process more trans­par­ent, great — though that’s eas­i­er said than done.

The point, rather, is that sim­plis­ti­cal­ly equat­ing less mod­er­a­tion with “free speech” — and more mod­er­a­tion with “cen­sor­ship” — doesn’t do any real-world work. There are all kinds of Twit­ter expres­sions where mod­er­a­tion isn’t an easy call — online harass­ment, rank dis­in­for­ma­tion, efforts to foment vio­lent insur­rec­tion or under­mine democ­ra­cy with delib­er­ate pro­pa­gan­da, and so on.

...

A vis­i­tor from, say, Mars might notice two big things com­ing from Repub­li­cans and right-wing pro­pa­gan­dists these days. First, they are wide­ly apply­ing the terms “groomer” and “pedophile” to Democ­rats and cor­po­ra­tions that oppose Repub­li­can bills restrict­ing class­room dis­cus­sion of sex and gen­der.

Sec­ond, they’re con­cert­ed­ly try­ing to impede efforts to get to the bot­tom of Don­ald Trump’s effort to over­turn the 2020 elec­tion. Mean­while, many Repub­li­cans are run­ning for office on undi­lut­ed pro­pa­gan­da about 2020, to install them­selves in posi­tions of con­trol over future elec­tions.

This stuff doesn’t map onto Musk’s dia­gram of left­ism and con­ser­vatism. As David Lurie notes, it’s not a move in a “con­ser­v­a­tive” direc­tion. It’s a turn away from the con­sti­tu­tion­al order itself. Musk is telling us that those abnor­mal­i­ties will not reg­is­ter in his under­stand­ing of the moment.

Yes, Musk denounces the “far right” along with the “far left.” But all this shows is that he’ll fall back on absurd equiv­a­lences to down­play how vir­u­lent and destruc­tive the for­mer has tru­ly become.

And notably, many mem­bers of the Repub­li­can Par­ty are traf­fick­ing in “groomer” smears and help­ing to feed the party’s insur­rec­tion­ist streak. As Bri­an Beut­ler shows, Repub­li­can offi­cials are threat­en­ing use of insti­tu­tion­al pow­er to help secure Musk’s Twit­ter takeover, mean­ing they see a big advan­tage in how that would alter the politi­co-infor­ma­tion envi­ron­ment.

What does this mean in prac­tice?

Danielle Cit­ron, a Uni­ver­si­ty of Vir­ginia law pro­fes­sor who stud­ies online harass­ment, sug­gests a worst-case sce­nario: A less-mod­er­at­ed Twit­ter might be weaponized in online harass­ment cam­paigns to paint Demo­c­ra­t­ic offi­cials — espe­cial­ly ones who attack GOP bills on sex and gen­der iden­ti­ty — as groomers and pedophiles.

“If Demo­c­ra­t­ic politi­cians are accused of being pedophiles,” Cit­ron tells me, you can imag­ine “deep fake sex videos” that employ depic­tions of child porn with the “politi­cian swapped in.”

Alter­na­tive­ly, says Cit­ron, you can see a much more exten­sive out­pour­ing of “dis­in­for­ma­tion” about our elec­tions, par­tic­u­lar­ly amid anoth­er coup attempt.

“Twit­ter looked like that 12 years ago,” Cit­ron tells me. Indeed, some cur­rent Twit­ter employ­ees fear return­ing to exact­ly that.

...

It may be that Musk might not end up allow­ing any­thing like this to hap­pen, once his vague “free speech” bro­mides col­lide with messy mod­er­a­tion real­i­ties. But when he dis­plays his deter­mi­na­tion to down­play the rad­i­cal­iza­tion of the right wing of the GOP, he’s show­ing us a poten­tial future infor­ma­tion land­scape that far-right Repub­li­cans are sure­ly dream­ing about.

No won­der they’re so excit­ed.

———–

“Elon Musk reveals how his Twit­ter may fuel right-wing extrem­ism” by Greg Sar­gent; The Wash­ing­ton Post; 04/29/2022

“It may be that Musk might not end up allow­ing any­thing like this to hap­pen, once his vague “free speech” bro­mides col­lide with messy mod­er­a­tion real­i­ties. But when he dis­plays his deter­mi­na­tion to down­play the rad­i­cal­iza­tion of the right wing of the GOP, he’s show­ing us a poten­tial future infor­ma­tion land­scape that far-right Repub­li­cans are sure­ly dream­ing about.”

The flood­gates are being opened. Which means it’s just a mat­ter of time before the worst kind of dis­in­for­ma­tion is once again flood­ing that plat­form. But it’s not just going to be a return to the bad old days of yes­ter­year. Deep Fake tech­nolo­gies did­n’t exist back when Twit­ter was last a free-for-all far right play­ground. It’s a brave new world. There’s more than one way to release a Krak­en:

...
What does this mean in prac­tice?

Danielle Cit­ron, a Uni­ver­si­ty of Vir­ginia law pro­fes­sor who stud­ies online harass­ment, sug­gests a worst-case sce­nario: A less-mod­er­at­ed Twit­ter might be weaponized in online harass­ment cam­paigns to paint Demo­c­ra­t­ic offi­cials — espe­cial­ly ones who attack GOP bills on sex and gen­der iden­ti­ty — as groomers and pedophiles.

“If Demo­c­ra­t­ic politi­cians are accused of being pedophiles,” Cit­ron tells me, you can imag­ine “deep fake sex videos” that employ depic­tions of child porn with the “politi­cian swapped in.”

Alter­na­tive­ly, says Cit­ron, you can see a much more exten­sive out­pour­ing of “dis­in­for­ma­tion” about our elec­tions, par­tic­u­lar­ly amid anoth­er coup attempt.

“Twit­ter looked like that 12 years ago,” Cit­ron tells me. Indeed, some cur­rent Twit­ter employ­ees fear return­ing to exact­ly that.
...

So Musk is com­ing out as a Repub­li­can at the same time he’s mak­ing this pur­chase of Twit­ter seem­ing­ly in oppo­si­tion to lefty ‘wokeism’. It cer­tain­ly gives us a major hint as to what to expect from Musk, at least when it comes to dis­in­for­ma­tion in US pol­i­tics. But how about Twit­ters oth­er prob­lem area when it comes to mod­er­a­tion: the over­mod­er­a­tion of any­thing involv­ing Chi­na or Rus­sia that does­n’t fit with the pre­vail­ing nar­ra­tives com­ing out of the US nation­al secu­ri­ty state? Can we at least expect some improve­ments there? Sure, if you believe some­one who is anx­ious­ly court­ing more and more Pen­ta­gon con­tracts is going to do any­thing to piss off his biggest cus­tomer:

Yasha Levine Sub­stack

Elon Musk is a spy

Yasha Levine
Apr 25, 2022

The lat­est out­rage to hit the wire is that the board of Twit­ter has agreed to accept Elon Musk’s buy­out pro­pos­al, which will give Elon full con­trol of the com­pa­ny and allow him to take it pri­vate. There’s lots of howl­ing all around, as if they sky is falling. If you ask me, Twit­ter — and social media in gen­er­al — is garbage tech that most­ly wastes our time and poi­sons our minds. But regard­less of where you stand sale of Twit­ter to a Twit­ter-addict­ed oli­garch, to me the deal just fur­ther proves the the­sis of my book, Sur­veil­lance Val­ley. The Inter­net is an exten­sion of the Amer­i­can Empire.

I mean here you have Elon — an “out­sider” — mount­ing a hos­tile takeover of a major glob­al com­mu­ni­ca­tion plat­form. And the thing about him is that he’s not just a suc­cess­ful lithi­um bat­tery sales­man, he’s also a major mil­i­tary con­trac­tor doing busi­ness with the most secre­tive and “strate­gi­cal­ly impor­tant” spooks in Amer­i­ca.

For exam­ple: Even as he was plan­ning his takeover this April, this was hap­pen­ing: the cul­mi­na­tion of a near­ly $300 mil­lion mil­i­tary con­tract to launch a clas­si­fied Amer­i­can spy satel­lite.

VANDENBERG SPACE FORCE BASE, Calif. (AP) — A clas­si­fied satel­lite for the U.S. Nation­al Recon­nais­sance Office was launched into space from Cal­i­for­nia on Sun­day.

The NROL-85 satel­lite lift­ed off at 6:13 a.m. from Van­den­berg Space Force Base aboard a two-stage SpaceX Fal­con 9 rock­et.

It was the first mis­sion by the NRO to reuse a SpaceX rock­et boost­er, Van­den­berg said in a state­ment.

…The NRO only described the NROL-85 satel­lite as a “crit­i­cal nation­al secu­ri­ty pay­load.”

Its launch was one of three award­ed by the Air Force to SpaceX in 2019 for a com­bined fixed price of $297 mil­lion.

The NRO is the gov­ern­ment agency in charge of devel­op­ing, build­ing, launch­ing and main­tain­ing U.S. satel­lites that pro­vide intel­li­gence data to senior pol­i­cy­mak­ers, the intel­li­gence com­mu­ni­ty and the Defense Depart­ment.

I haven’t done a full account­ing of Elon Musk’s mil­i­tary con­trac­tor his­to­ry, but this is not a one-off thing for him. Since going into the space busi­ness, he’s angled to be a deal­er of pri­vate rock­et and satel­lite ser­vices to the Pen­ta­gon. In 2019, he got fund­ing to test his Star­link satel­lites so they could route encrypt­ed coms for the Air Force. 2020, he signed a $149 mil­lion deal to track mis­siles — aka to spy on the sky. And of course, in 2021, he won a $2.89 bil­lion con­tract with NASA. And I’m sure there is much much more there — I just haven’t been pay­ing much atten­tion to the nexus of Sil­i­con Val­ley and America’s sprawl­ing secu­ri­ty state over the last few years, as I’ve moved on to oth­er things after writ­ing Sur­veil­lance Val­ley.

...

———-

“Elon Musk is a spy” by Yasha Levine; Yasha Levine Sub­stack; 04/25/2022

“I mean here you have Elon — an “out­sider” — mount­ing a hos­tile takeover of a major glob­al com­mu­ni­ca­tion plat­form. And the thing about him is that he’s not just a suc­cess­ful lithi­um bat­tery sales­man, he’s also a major mil­i­tary con­trac­tor doing busi­ness with the most secre­tive and “strate­gi­cal­ly impor­tant” spooks in Amer­i­ca.”

Musk is clear­ly more than hap­py to piss of ‘the left’. He’s kind of mak­ing that his per­son­al brand at this point. But how about the Pen­ta­gon? As the con­ser­v­a­tive stick fig­ure in Musk’s tweet put it, LOL!

Fol­low­ing up on the uproar over Elon Musk’s pur­chase of Twit­ter and, as Yasha Levine point­ed out, the com­plete lack of any acknowl­edge­ment in that uproar over Musk’s grow­ing sta­tus as a major US nation­al secu­ri­ty con­trac­tor, here’s a post on the Law­fare blog from last month that under­scores anoth­er aspect of Musk’s rela­tion­ship with the US nation­al secu­ri­ty state: the dual use nature of Musk’s Star­link satel­lites net­work and the fact that it’s already being used for mil­i­tary pur­pos­es. In Ukraine. Yep, it turns out Musk’s Star­link satel­lite net­work has been play­ing a cru­cial role in pro­vid­ing inter­net con­nec­tiv­i­ty for Ukraine’s mil­i­tary. A role that was encour­aged by USAID. In fact, USAID issued a press release last month tout­ing how it set up a pub­lic-pri­vate part­ner­ship with Star­link to send 5000 Star­link ter­mi­nals to Ukraine to main­tain inter­net con­nec­tiv­i­ty dur­ing the war. And as the fol­low­ing Law­fare blog post points out, that use has­n’t been lim­it­ed to civil­ian uses. One Ukrain­ian com­man­der told the Times of Lon­don that they “must” use Star­link to tar­get Russ­ian sol­diers at night with ther­mal imag­ing.

So Musk deliv­ered a large num­ber of Star­link ter­mi­nals to Ukraine under a USAID pro­gram to pro­vide civil­ians with inter­net con­nec­tiv­i­ty and they end up get­ting used by Ukraine’s mil­i­tary. It’s the kind of sit­u­a­tion that cre­ates a num­ber of pos­si­ble legal headaches. As we’re going to see, the US Space Com­mand has already set up a pro­gram for incor­po­rat­ing com­mer­cial infra­struc­ture oper­at­ing in space into mil­i­tary efforts and these Star­link satel­lites are read­i­ly capa­ble of han­dling the Com­mand, Con­trol, Com­mu­ni­ca­tions, Com­put­ers, Intel­li­gence, Sur­veil­lance and Recon­nais­sance (C4ISR) func­tions nec­es­sary for mod­ern mil­i­tary oper­a­tions.

But per­haps the biggest pos­si­ble headache that could emerge from this is the one experts have been warn­ing us about ever since Musk hatched this Star­link scheme: the threat of a space junk cas­cade that makes the earth­’s low orbit space effec­tive­ly unwork­able. That kind of sce­nario was already a risk just from things going wrong. And now we’re learn­ing that Musk is allow­ing Star­link to be used for exact­ly the kind of activ­i­ties that could prompt a phys­i­cal attack on the Star­link clus­ter:

Law­fare Blog

What Does Starlink’s Par­tic­i­pa­tion in Ukrain­ian Defense Reveal About U.S. Space Pol­i­cy?

By Matthew Fitzger­ald, Cort Thomp­son
Tues­day, April 26, 2022, 8:01 AM

Recent sig­nal jam­ming and cyber­at­tacks have bat­tered Ukrain­ian dig­i­tal net­work infra­struc­ture and caused severe out­ages. Reports attribute these attacks, which are con­sis­tent with pri­or Rus­sia-spon­sored oper­a­tions, to Russ­ian actors.

In an effort to secure redun­dant com­mu­ni­ca­tions capa­bil­i­ties, Ukrain­ian Min­is­ter of Dig­i­tal Trans­for­ma­tion Mykhai­lo Federov appealed to Elon Musk for Star­link-enabled inter­net. In response, Musk’s com­pa­ny SpaceX acti­vat­ed Star­link ser­vice in Ukraine and sent addi­tion­al net­work ter­mi­nals. And while USAID posits that the ter­mi­nals are nec­es­sary to “safe­guard” pub­lic inter­net access, Star­link has report­ed­ly enabled mem­bers of the Ukrain­ian Aero­rozvid­ka to car­ry out sophis­ti­cat­ed intel­li­gence col­lec­tion and fire sup­port oper­a­tions against Russ­ian posi­tions. One Ukrain­ian mil­i­tary offi­cial even told the Times of Lon­don that he “must” use Star­link in order to acquire Russ­ian tar­gets with ther­mal imag­ing.

Musk cor­re­sponds with the Ukrain­ian gov­ern­ment against the back­drop of a com­plex legal land­scape. This post explores sev­er­al tenets of inter­na­tion­al human­i­tar­i­an law as it might gov­ern Russ­ian tar­get­ing of Star­link infra­struc­ture. It then assess­es how and why Musk’s actions threat­en to draw the U.S. in as a par­ty to the con­flict. Final­ly, it pro­pos­es mod­i­fi­ca­tions to domes­tic pol­i­cy that could help avoid such an out­come now and in the future.

Inter­na­tion­al Human­i­tar­i­an Law and Dual-Use Objects

Inter­na­tion­al human­i­tar­i­an law pro­vides a frame­work for iden­ti­fy­ing law­ful tar­gets in war. Although states may dis­agree in their inter­pre­ta­tions of spe­cif­ic treaty pro­vi­sions, most agree on cer­tain foun­da­tion­al prin­ci­ples.

First, com­bat­ants have a duty to dis­tin­guish between tar­gets that con­sti­tute mil­i­tary objec­tives and those that do not. Mil­i­tary objec­tives include only those com­bat­ants or objects that, based on their nature, loca­tion, pur­pose, or use, make an effec­tive con­tri­bu­tion to mil­i­tary action. Civil­ians enjoy pro­tec­tion under this pro­vi­sion “unless and for such time” as they might sub­stan­tial­ly and effec­tive­ly enhance or sus­tain a war­mak­ing effort. Com­bat­ants include uni­formed mil­i­tary per­son­nel belong­ing to a par­ty to a con­flict, as well as nonuni­formed per­son­nel over which a par­ty exer­cis­es over­all con­trol.

Sec­ond, com­bat­ants must avoid car­ry­ing out any strike that may cause suf­fer­ing, injury, or destruc­tion to non­com­bat­ants or civil­ians objects that would be clear­ly exces­sive in rela­tion to the con­crete and direct over­all mil­i­tary advan­tage antic­i­pat­ed. In inter­na­tion­al human­i­tar­i­an law, this is known as the prin­ci­ple of pro­por­tion­al­i­ty.

Final­ly, attack­ers must take all fea­si­ble mea­sures to pre­vent col­lat­er­al dam­age to non­mil­i­tary objec­tives, includ­ing a duty to warn civil­ian pop­u­la­tions that might be affect­ed. This oblig­a­tion applies only where cir­cum­stances per­mit such warn­ing.

Dual-use objects, or objects that serve both a mil­i­tary and a civil­ian use, pro­vide unique chal­lenges to mil­i­tary com­man­ders. Dual-use objects may be tar­gets of attack only dur­ing those times when they are used for a mil­i­tary pur­pose, and when such a strike would also not be dis­pro­por­tion­ate. As sys­tem ele­ments become more atten­u­at­ed from the mil­i­tary actions they enable, and increas­ing­ly com­min­gled with non­mil­i­tary pur­pos­es, pro­por­tion­al­i­ty and mil­i­tary neces­si­ty cal­cu­lus tends to dis­fa­vor attack. How­ev­er, each case-by-case deter­mi­na­tion depends on the infor­ma­tion rea­son­ably avail­able to mil­i­tary com­man­ders and the tem­po­ral­i­ty ele­ment regard­ing a par­tic­u­lar object’s use.

Dual-use objects that may at times be sub­ject to attack include sev­er­al relics of tra­di­tion­al bat­tle­fields, such as oil refiner­ies, bridges, tun­nels, trans­mis­sion facil­i­ties and fac­to­ries. How­ev­er, the nature of war has expand­ed to addi­tion­al and incom­pa­ra­ble bat­tle­field dimen­sions. For exam­ple, future engage­ments in the cyber or space domains will like­ly require com­man­ders to make real-time deter­mi­na­tions on tar­gets expe­ri­enc­ing con­cur­rent mil­i­tary and civil­ian use. Lega­cy rules for tar­get­ing dual-use infra­struc­ture may not ade­quate­ly address these new dimen­sions.

Are Star­link Satel­lites Now Law­ful Tar­gets?

In an age in which Com­mand, Con­trol, Com­mu­ni­ca­tions, Com­put­ers, Intel­li­gence, Sur­veil­lance and Recon­nais­sance (C4ISR) sys­tems are essen­tial to enable mil­i­tary oper­a­tions, the satel­lites that per­form C4ISR func­tions almost cer­tain­ly offer a mil­i­tary advan­tage. By Musk’s own assess­ment, the “prob­a­bil­i­ty of (Star­link com­po­nents) being tar­get­ed is high.” While C4ISR sys­tems may be civil­ian owned and oper­at­ed (such as Star­link), their inte­gra­tion into mil­i­tary net­works sug­gests that tem­po­ral “use” analy­ses may be sat­is­fied dur­ing peri­ods of armed con­flict. Although a denial of ser­vice or oth­er adverse effect levied on a Star­link satel­lite would be dif­fi­cult to local­ize and would inevitably cause harm to civil­ian infra­struc­ture, the dis­tinc­tion prong of a tar­get­ing analy­sis might still be sat­is­fied.

Achiev­ing pro­por­tion­al­i­ty when attack­ing satel­lites in orbit pos­es unique chal­lenges. First, at the same time that Starlink’s satel­lite net­work enables Ukrain­ian mil­i­tary offi­cials to achieve a mil­i­tary advan­tage, it also pro­vides ser­vice to hun­dreds of thou­sands of civil­ian users. Rus­sia might pos­sess only lim­it­ed capac­i­ty to local­ize denials of ser­vice or tai­lor effects to counter only mil­i­tary uses. Sec­ond, unlike kinet­ic effects at sea lev­el, attacks that cause frag­men­ta­tion in space levy inevitable but unpre­dictable effects on oth­er non­mil­i­tary struc­tures. Although adverse effects on the envi­ron­ment should weigh heav­i­ly against attack, they are not yet a per se bar to pur­su­ing kinet­ic effects in out­er space.

Despite unique dif­fi­cul­ties in sat­is­fy­ing prin­ci­ples of inter­na­tion­al human­i­tar­i­an law while tar­get­ing satel­lites in orbit, Rus­sia would argue that degrad­ing one of the Ukrain­ian military’s most viable com­mu­ni­ca­tions sys­tems offers exhor­bi­tant mil­i­tary val­ue. Fur­ther­more, Russ­ian com­man­ders might claim that addi­tion­al pre­cau­tions are not fea­si­ble, that Russ­ian Pres­i­dent Vladimir Putin’s assess­ment of Ukrain­ian war crimes jus­ti­fies esca­la­tion in a new domain, or oth­er alle­ga­tions not nec­es­sar­i­ly root­ed in fact nor law.

Per­haps even more con­cern­ing, Rus­sia has demon­strat­ed a a will­ing­ness to intro­duce entropy in the space domain in order to advance nation­al inter­ests with lit­tle regard for notions of accept­able con­duct. Addi­tion­al inci­dents of sim­i­lar con­duct should not come as a sur­prise, espe­cial­ly where Rus­sia is already alleged to have launched indis­crim­i­nate and dis­pro­por­tion­ate attacks dur­ing its advance toward Kyiv. In light of these trends, the Unit­ed States must be pre­pared to counter Russ­ian actions against nation­al inter­ests, which include pro­tect­ing the domes­tic com­mer­cial space sec­tor and its sup­port­ing infra­struc­ture.

Does the Unit­ed States Bear the Risk of SpaceX’s Actions?

Star­link satel­lites are the prop­er­ty of SpaceX and its share­hold­ers, but the U.S. also has a vest­ed inter­est in gov­ern­ing their con­tri­bu­tions to for­eign mil­i­tary oper­a­tions. Nation­al Space Pol­i­cy (NSP) directs the sec­re­tary of com­merce to deter­mine whether nov­el com­mer­cial activ­i­ties threat­en inter­na­tion­al oblig­a­tions. Starlink’s use like­ly falls in a gap between the NSP’s export poli­cies for tech­nol­o­gy and the antic­i­pat­ed areas of super­vi­sion. After all, the satel­lite con­stel­la­tion itself is mar­ket­ed for com­mer­cial use, and the only exports are user ter­mi­nals enabling con­nec­tiv­i­ty.

This dis­tinc­tion may abdi­cate respon­si­bil­i­ty under an over­all con­trol analy­sis, but the Out­er Space Treaty of 1967 (OST) offers no such safe har­bor. OST Arti­cle VI attach­es state respon­si­bil­i­ty to activ­i­ties in out­er space, includ­ing those launched from a state’s ter­ri­to­ry, even when the activ­i­ty is car­ried out by a non­govern­men­tal agency. More specif­i­cal­ly, Arti­cle VI demands that “the activ­i­ties of non-gov­ern­men­tal enti­ties in out­er space … require autho­riza­tion and con­tin­u­ing super­vi­sion by the appro­pri­ate State Par­ty” (empha­sis added).

The USSR first pro­posed assign­ment of nation­al respon­si­bil­i­ty for non­govern­ment enti­ties’ space activ­i­ties in 1962. It is quite pos­si­ble the draft treaty amend­ment sub­mit­ted on July 11, 1966, and the result­ing text of the OST, can be attrib­uted to delib­er­a­tive pro­gram­ming aimed at address­ing the ide­o­log­i­cal com­pe­ti­tion between free-mar­ket and cen­tral­ly planned eco­nom­ic actors. Now, the ready accep­tance of state attri­bu­tion by the U.S. del­e­ga­tion may impute con­se­quences for Musk’s con­tri­bu­tion to the cur­rent con­flict. This should cause con­cern for pol­i­cy­mak­ers who are weigh­ing future con­tri­bu­tions of arms and equip­ment to Ukraine because no domes­tic leg­isla­tive mech­a­nism has been iden­ti­fied to autho­rize, or to restrain, Starlink’s cur­rent wartime con­tri­bu­tions.

Though it may be too late to recall Musk’s Star­link ter­mi­nals, the Sen­ate Armed Ser­vices Com­mit­tee is active­ly con­sid­er­ing the impli­ca­tions of pri­vate actors in space. U.S. Space Com­mand also acknowl­edges hav­ing explored the legal frame­work for com­mer­cial assis­tance in armed con­flicts in coor­di­na­tion with the Com­mer­cial Inte­gra­tion Cell (CIC). Accord­ing to U.S. Strate­gic Com­mand, the CIC is the “first-ever col­lab­o­ra­tive gov­ern­ment and indus­try effort to inte­grate com­mer­cial satel­lite owner/operators into the Com­bined Space Oper­a­tions Cen­ter.” Depend­ing on use cas­es CIC par­tic­i­pants agree to, the com­mer­cial assets may cross the thresh­old of dual-use infra­struc­ture that pro­vide both mil­i­tary and civil­ian func­tion­al­i­ty. In these cas­es, it is like­ly increas­ing­ly dif­fi­cult to deter­mine a moment when an object los­es its civil­ian sta­tus pro­tec­tion due to simul­ta­ne­ous use.

The CIC appears primed to address coor­di­na­tion of cer­tain activ­i­ties in out­er space, but revi­sions to the NSP are nec­es­sary to address the unan­tic­i­pat­ed con­se­quences of Arti­cle VI respon­si­bil­i­ties. While the NSP clear­ly artic­u­lates the ram­i­fi­ca­tions of an attack on U.S. space infra­struc­ture, Starlink’s activ­i­ties are invit­ing exact­ly such a threat. Unless the U.S. intends to take a per­mis­sive stance toward real or per­ceived cor­po­rate insti­ga­tion in out­er space, and is pre­pared to enter armed con­flicts accord­ing­ly, indus­try coop­er­a­tion alone is insuf­fi­cient. Dual-use objects have devel­oped into con­cur­rent-use objects, and nation­al pol­i­cy must care­ful­ly ana­lyze and super­vise for­eign pro­vi­sion of materiel in the space domain.

...

Dual-use space objects are increas­ing­ly trans­form­ing into con­cur­rent-use infra­struc­ture, obscur­ing defined peri­ods of pro­tect­ed sta­tus, and expos­ing U.S. assets to a more per­mis­sive tar­get­ing cal­cu­lus. Fur­ther, Depart­ment of Com­merce review and super­vi­sion of nov­el com­mer­cial activ­i­ties has proved insuf­fi­cient to pre­vent unin­tend­ed par­tic­i­pa­tion in for­eign con­flicts. The next NSP should man­date a delib­er­ate review process for pos­si­ble Arti­cle VI ram­i­fi­ca­tions pri­or to com­mer­cial actors pro­vid­ing any materiel that might impute state respon­si­bil­i­ty. In the mean­time, rein­forc­ing the reper­cus­sions for tar­get­ing any U.S.-flagged space asset is like­ly nec­es­sary to pre­vent Russ­ian tar­get­ing beyond ground-based ter­mi­nals. Oth­er­wise, Musk’s irrev­er­ent taunt­ing may land uncom­fort­ably close to a dec­la­ra­tion of war.

———-

“What Does Starlink’s Par­tic­i­pa­tion in Ukrain­ian Defense Reveal About U.S. Space Pol­i­cy?” by Matthew Fitzger­ald, Cort Thomp­son; Law­fare Blog; 04/26/2022

“Musk cor­re­sponds with the Ukrain­ian gov­ern­ment against the back­drop of a com­plex legal land­scape. This post explores sev­er­al tenets of inter­na­tion­al human­i­tar­i­an law as it might gov­ern Russ­ian tar­get­ing of Star­link infra­struc­ture. It then assess­es how and why Musk’s actions threat­en to draw the U.S. in as a par­ty to the con­flict. Final­ly, it pro­pos­es mod­i­fi­ca­tions to domes­tic pol­i­cy that could help avoid such an out­come now and in the future.”

What are the impli­ca­tions of Elon Musk’s Star­link satel­lites being used by the Ukrain­ian mil­i­tary? Well, for starters, those Star­link satel­lites — which Musk has admit­ted are capa­ble of exe­cut­ing the Com­mand, Con­trol, Com­mu­ni­ca­tions, Com­put­ers, Intel­li­gence, Sur­veil­lance and Recon­nais­sance (C4ISR) func­tions required by mod­ern mil­i­taries — are clear­ly “dual use” pieces of infra­struc­ture. And under inter­na­tion­al law that means these satel­lites could poten­tial­ly be legal­ly attacked by Rus­sia. So one very direct impli­ca­tion of the Ukrain­ian mil­i­tary’s use of Musk’s net­work is a pos­si­ble Russ­ian attack on a com­meri­cial satel­lite sys­tem oper­at­ed by a US com­pa­ny:

...
In an effort to secure redun­dant com­mu­ni­ca­tions capa­bil­i­ties, Ukrain­ian Min­is­ter of Dig­i­tal Trans­for­ma­tion Mykhai­lo Federov appealed to Elon Musk for Star­link-enabled inter­net. In response, Musk’s com­pa­ny SpaceX acti­vat­ed Star­link ser­vice in Ukraine and sent addi­tion­al net­work ter­mi­nals. And while USAID posits that the ter­mi­nals are nec­es­sary to “safe­guard” pub­lic inter­net access, Star­link has report­ed­ly enabled mem­bers of the Ukrain­ian Aero­rozvid­ka to car­ry out sophis­ti­cat­ed intel­li­gence col­lec­tion and fire sup­port oper­a­tions against Russ­ian posi­tions. One Ukrain­ian mil­i­tary offi­cial even told the Times of Lon­don that he “must” use Star­link in order to acquire Russ­ian tar­gets with ther­mal imag­ing.

...

Dual-use objects, or objects that serve both a mil­i­tary and a civil­ian use, pro­vide unique chal­lenges to mil­i­tary com­man­ders. Dual-use objects may be tar­gets of attack only dur­ing those times when they are used for a mil­i­tary pur­pose, and when such a strike would also not be dis­pro­por­tion­ate. As sys­tem ele­ments become more atten­u­at­ed from the mil­i­tary actions they enable, and increas­ing­ly com­min­gled with non­mil­i­tary pur­pos­es, pro­por­tion­al­i­ty and mil­i­tary neces­si­ty cal­cu­lus tends to dis­fa­vor attack. How­ev­er, each case-by-case deter­mi­na­tion depends on the infor­ma­tion rea­son­ably avail­able to mil­i­tary com­man­ders and the tem­po­ral­i­ty ele­ment regard­ing a par­tic­u­lar object’s use.

Dual-use objects that may at times be sub­ject to attack include sev­er­al relics of tra­di­tion­al bat­tle­fields, such as oil refiner­ies, bridges, tun­nels, trans­mis­sion facil­i­ties and fac­to­ries. How­ev­er, the nature of war has expand­ed to addi­tion­al and incom­pa­ra­ble bat­tle­field dimen­sions. For exam­ple, future engage­ments in the cyber or space domains will like­ly require com­man­ders to make real-time deter­mi­na­tions on tar­gets expe­ri­enc­ing con­cur­rent mil­i­tary and civil­ian use. Lega­cy rules for tar­get­ing dual-use infra­struc­ture may not ade­quate­ly address these new dimen­sions.

...

In an age in which Com­mand, Con­trol, Com­mu­ni­ca­tions, Com­put­ers, Intel­li­gence, Sur­veil­lance and Recon­nais­sance (C4ISR) sys­tems are essen­tial to enable mil­i­tary oper­a­tions, the satel­lites that per­form C4ISR func­tions almost cer­tain­ly offer a mil­i­tary advan­tage. By Musk’s own assess­ment, the “prob­a­bil­i­ty of (Star­link com­po­nents) being tar­get­ed is high.” While C4ISR sys­tems may be civil­ian owned and oper­at­ed (such as Star­link), their inte­gra­tion into mil­i­tary net­works sug­gests that tem­po­ral “use” analy­ses may be sat­is­fied dur­ing peri­ods of armed con­flict. Although a denial of ser­vice or oth­er adverse effect levied on a Star­link satel­lite would be dif­fi­cult to local­ize and would inevitably cause harm to civil­ian infra­struc­ture, the dis­tinc­tion prong of a tar­get­ing analy­sis might still be sat­is­fied.
...

And if Rus­sia does indeed decide to launch some sort of attack against Star­link, what can we expect the US to do in response? Well, accord­ing to Arti­cle VI of the Out­er Space Treaty of 1967, state respon­si­bil­i­ty to activ­i­ties in out­er space, includ­ing those launched from a state’s ter­ri­to­ry, even when the activ­i­ty is car­ried out by a non­govern­men­tal agency. Beyond that, the US Space Com­mand has already set up a Com­mer­cial Inte­gra­tion Cell (CIC) pro­gram designed to enlist the use of com­mer­cial satel­lite capa­bil­i­ties in armed con­flicts. So if there’s a Russ­ian attack on Star­link, it’s not nec­es­sar­i­ly going to be easy for the US to avoid esca­lat­ing the sit­u­a­tion because the US gov­ern­ment will be legal­ly respon­si­ble, in part, for sanc­tion­ing and coor­di­nat­ing the mil­i­tary use of this com­mer­cial infra­struc­ture. In oth­er words, a Russ­ian attack on Star­link could cre­ate a very messy sit­u­a­tion:

...
Star­link satel­lites are the prop­er­ty of SpaceX and its share­hold­ers, but the U.S. also has a vest­ed inter­est in gov­ern­ing their con­tri­bu­tions to for­eign mil­i­tary oper­a­tions. Nation­al Space Pol­i­cy (NSP) directs the sec­re­tary of com­merce to deter­mine whether nov­el com­mer­cial activ­i­ties threat­en inter­na­tion­al oblig­a­tions. Starlink’s use like­ly falls in a gap between the NSP’s export poli­cies for tech­nol­o­gy and the antic­i­pat­ed areas of super­vi­sion. After all, the satel­lite con­stel­la­tion itself is mar­ket­ed for com­mer­cial use, and the only exports are user ter­mi­nals enabling con­nec­tiv­i­ty.

This dis­tinc­tion may abdi­cate respon­si­bil­i­ty under an over­all con­trol analy­sis, but the Out­er Space Treaty of 1967 (OST) offers no such safe har­bor. OST Arti­cle VI attach­es state respon­si­bil­i­ty to activ­i­ties in out­er space, includ­ing those launched from a state’s ter­ri­to­ry, even when the activ­i­ty is car­ried out by a non­govern­men­tal agency. More specif­i­cal­ly, Arti­cle VI demands that “the activ­i­ties of non-gov­ern­men­tal enti­ties in out­er space … require autho­riza­tion and con­tin­u­ing super­vi­sion by the appro­pri­ate State Par­ty” (empha­sis added)

...

Though it may be too late to recall Musk’s Star­link ter­mi­nals, the Sen­ate Armed Ser­vices Com­mit­tee is active­ly con­sid­er­ing the impli­ca­tions of pri­vate actors in space. U.S. Space Com­mand also acknowl­edges hav­ing explored the legal frame­work for com­mer­cial assis­tance in armed con­flicts in coor­di­na­tion with the Com­mer­cial Inte­gra­tion Cell (CIC). Accord­ing to U.S. Strate­gic Com­mand, the CIC is the “first-ever col­lab­o­ra­tive gov­ern­ment and indus­try effort to inte­grate com­mer­cial satel­lite owner/operators into the Com­bined Space Oper­a­tions Cen­ter.” Depend­ing on use cas­es CIC par­tic­i­pants agree to, the com­mer­cial assets may cross the thresh­old of dual-use infra­struc­ture that pro­vide both mil­i­tary and civil­ian func­tion­al­i­ty. In these cas­es, it is like­ly increas­ing­ly dif­fi­cult to deter­mine a moment when an object los­es its civil­ian sta­tus pro­tec­tion due to simul­ta­ne­ous use.

The CIC appears primed to address coor­di­na­tion of cer­tain activ­i­ties in out­er space, but revi­sions to the NSP are nec­es­sary to address the unan­tic­i­pat­ed con­se­quences of Arti­cle VI respon­si­bil­i­ties. While the NSP clear­ly artic­u­lates the ram­i­fi­ca­tions of an attack on U.S. space infra­struc­ture, Starlink’s activ­i­ties are invit­ing exact­ly such a threat. Unless the U.S. intends to take a per­mis­sive stance toward real or per­ceived cor­po­rate insti­ga­tion in out­er space, and is pre­pared to enter armed con­flicts accord­ing­ly, indus­try coop­er­a­tion alone is insuf­fi­cient. Dual-use objects have devel­oped into con­cur­rent-use objects, and nation­al pol­i­cy must care­ful­ly ana­lyze and super­vise for­eign pro­vi­sion of materiel in the space domain.
...

But per­haps the biggest mess that could be cre­at­ed by a Russ­ian attack on Star­lin would be the lit­er­al mess in space that could result from any kinet­ic attacks on those satel­lites. As we’ve seen, Star­link already pos­es an unprece­dent­ed threat to the earth­’s orbital space, with a worst case sce­nario that could lit­ter the space around the plan­et with so much space junk it’s effec­tive­ly impos­si­ble to launch new objects into orbit. And as this Law­fare post notes, while the risk of cre­at­ing a bunch of space junk could cer­tain­ly give Rus­sia pause when con­sid­er­ing whether or not to car­ry out a kinet­ic attack on those satel­lites, that risk does­n’t bar Rus­sia from car­ry­ing out such an attack. There’s no inter­na­tion­al law against it so it’s real­ly up to Rus­sia at that moment to weigh the costs and ben­e­fits. So if these deci­sions are being made dur­ing a time when Ukraine’s mil­i­tary is evis­cer­at­ing Russ­ian forces in part on the use of those satel­lites, don’t be shocked if Rus­si­a’s cost/benefit analy­sis does­n’t leave clean plan­e­tary orbits at the top of the pri­or­i­ty list:

...
Achiev­ing pro­por­tion­al­i­ty when attack­ing satel­lites in orbit pos­es unique chal­lenges. First, at the same time that Starlink’s satel­lite net­work enables Ukrain­ian mil­i­tary offi­cials to achieve a mil­i­tary advan­tage, it also pro­vides ser­vice to hun­dreds of thou­sands of civil­ian users. Rus­sia might pos­sess only lim­it­ed capac­i­ty to local­ize denials of ser­vice or tai­lor effects to counter only mil­i­tary uses. Sec­ond, unlike kinet­ic effects at sea lev­el, attacks that cause frag­men­ta­tion in space levy inevitable but unpre­dictable effects on oth­er non­mil­i­tary struc­tures. Although adverse effects on the envi­ron­ment should weigh heav­i­ly against attack, they are not yet a per se bar to pur­su­ing kinet­ic effects in out­er space.

Despite unique dif­fi­cul­ties in sat­is­fy­ing prin­ci­ples of inter­na­tion­al human­i­tar­i­an law while tar­get­ing satel­lites in orbit, Rus­sia would argue that degrad­ing one of the Ukrain­ian military’s most viable com­mu­ni­ca­tions sys­tems offers exhor­bi­tant mil­i­tary val­ue. Fur­ther­more, Russ­ian com­man­ders might claim that addi­tion­al pre­cau­tions are not fea­si­ble, that Russ­ian Pres­i­dent Vladimir Putin’s assess­ment of Ukrain­ian war crimes jus­ti­fies esca­la­tion in a new domain, or oth­er alle­ga­tions not nec­es­sar­i­ly root­ed in fact nor law.
...

Also keep in mind that the unique vul­ner­a­bil­i­ty the Star­link sys­tem — cre­at­ing a night­mare space junk cas­cade due to the large num­ber of low orbit tiny satel­lites Musk just launched into orbit with­out any seri­ous con­sid­er­a­tion of the risks — is exact­ly the kind of thing oppos­ing mil­i­taries might be tempt­ed to cre­ate as part of a big gam­ble. Who would be hurt more by a space junk cas­cade that crip­ples com­mer­cial space activ­i­ty? An already eco­nom­i­cal­ly crip­pled Rus­sia, or the US? It would­n’t nec­es­sar­i­ly take the frag­men­ta­tion of that many of these satel­lites to get a cas­cade start­ed. Or maybe they’ll just threat­en to do it. Either way, let’s hope the Russ­ian gov­ern­ment, and any oth­er gov­ern­ments direct­ly threat­ened by Star­link in the future con­flicts, are actu­al­ly tak­ing these risks seri­ous­ly because it’s obvi­ous the peo­ple deploy­ing and using the sys­tem are not.

Who is going to pre­vent Elon Musk’s Star­link net­work of microsatel­lites from turn­ing the earth­’s low­er orbits into a swarm of lethal space junk that threat­ens to inca­pac­i­tate our abil­i­ty to oper­ate in space? No one, prob­a­bly. That’s the like­ly answer we can infer from the fol­low­ing pairs of arti­cles about Star­link.

The first arti­cle, from last month, high­lights a rather inter­est­ing anom­aly observed between Star­link and USAID. As we’ve seen, USAID cre­at­ed some sort of pub­lic-pri­vate part­ner­ship with Star­link for the deliv­ery of 5,000 Star­link ter­mi­nals to Ukraine to help deliv­er inter­net ser­vices to the coun­try. Includ­ing vital inter­net ser­vices for Ukraine’s mil­i­tary, rais­ing obvi­ous­ly ques­tions about whether or not Star­link could poten­tial­ly come under attack by Rus­sia.

How much mon­ey did USAID pro­vide for this ini­tia­tive? Well, that’s part of the mys­tery. The oth­er part of the mys­tery is what exact­ly did USAID pay for. We’re told that USAID paid SpaceX $1,500 per Star­link ter­mi­nal for 1,333 ter­mi­nals, adding up to $2 mil­lion. The stan­dard Star­link ter­mi­nal costs $600 while there’s a more advanced ver­sion that sells for $2,500. So was USAID pay­ing $1,500 for the $600 ter­mi­nals? If so, that would be some rather out­ra­geous price goug­ing, so maybe it was $1,500 for the $2,500 ter­mi­nals. We don’t know, but adding to the mys­tery is that USAID altered its pub­lic state­ments on this pub­lic-pri­vate part­ner­ship. The ini­tial April 5 state­ment released by USAID not­ed that SpaceX donat­ed 3,667 ter­mi­nals while USAID pur­chased an addi­tion­al 1,333 ter­mi­nals. Those num­bers were removed from an update released lat­er that day.

So for what­ev­er rea­son, USAID behaved in a way that sug­gest­ed some degree of sen­si­tiv­i­ty about these num­bers. We don’t know why, but what is clear from this sto­ry is that the US gov­ern­ment sees a lot of val­ue in Star­link’s capa­bil­i­ties. Which is rather prob­lem­at­ic when it comes to reg­u­lat­ing Star­link and ensur­ing it does­n’t pose an unrea­son­able risk of a space junk cas­cade cat­a­stro­phe. And that brings us to the sec­ond arti­cle below from back in August of last year. The sto­ry is about a study done by researchers on the rate of orbital close encoun­ters since the launch of Star­link. Basi­cal­ly, they’ve dou­bled in the last cou­ple of years, with half of the close encoun­ters involv­ing Star­link satel­lites. So Star­link is already prov­ing itself to be a major space col­li­sion haz­ard, and it’s bare­ly even fin­ished yet. Yes, as of the time of that arti­cle, only 1,700 Star­link satel­lites were in orbit. The plan is for tens of thou­sands of them to even­tu­al­ly be launched into orbit. That’s why these researchers were pre­dict­ing that 90% of orbital close encoun­ters in the future are like­ly to involve Star­link satel­lites.

And that’s why the mys­tery regard­ing Star­link’s rela­tion­ship to USAID, and SpaceX’s larg­er rela­tion­ship to the US nation­al secu­ri­ty state, could end up being a rather cru­cial ques­tion in terms of whether or not any­thing is going to be done to pre­vent an orbital space junk cas­cade cat­a­stro­phe. Because it sure does­n’t look like the US gov­ern­ment is over­ly con­cerned with these risks right now. Quite the oppo­site:

The Verge

The US gov­ern­ment report­ed­ly paid to send Star­link ter­mi­nals to Ukraine

SpaceX pre­vi­ous­ly said the US didn’t pro­vide finan­cial aid

By Mitchell Clark
Apr 8, 2022, 2:01pm EDT

Despite SpaceX imply­ing that the US didn’t give mon­ey to send Star­link ter­mi­nals to Ukraine in March, a report from The Wash­ing­ton Post reveals that the gov­ern­ment actu­al­ly paid mil­lions of dol­lars for equip­ment and trans­porta­tion. The report found that the US Agency for Inter­na­tion­al Devel­op­ment, or USAID, paid $1,500 apiece for 1,333 ter­mi­nals, adding up to around $2 mil­lion. USAID dis­closed the num­ber of ter­mi­nals it bought from the com­pa­ny in a press release from ear­ly April that has since been altered to remove men­tions of the pur­chase.

Accord­ing to space reporter Joey Roulette, SpaceX donat­ed 3,667 ter­mi­nals to Ukraine, or around $10 mil­lion worth, after also fac­tor­ing in the three months of data it pro­vid­ed with the ter­mi­nals. How­ev­er, it’s pos­si­ble this dona­tion was par­tial­ly sub­si­dized by USAID’s $1,500-a-unit pur­chase. The Wash­ing­ton Post says it’s unclear whether Ukraine received stan­dard ter­mi­nals, which SpaceX cur­rent­ly charges $600 for, or the advanced ter­mi­nals, which were announced in Feb­ru­ary and cost $2,500.

Looks like USAID edit­ed their press release and delet­ed the val­ue of SpaceX’s dona­tion, as well as the num­ber of ter­mi­nals the agency said SpaceX pro­vid­ed on its own dime. Here’s the pre­vi­ous ver­sion vs. cur­rent ver­sion: https://t.co/ws1urZu4VM pic.twitter.com/1eZxfQ6DPA— Joey Roulette (@joroulette) April 6, 2022

If USAID pur­chased reg­u­lar ter­mi­nals, it paid $900 over retail cost per unit (adding up to around $1.2 mil­lion) — notably, a recent report said that the ter­mi­nals cost SpaceX around $1,000 to build. It is pos­si­ble that USAID was also pay­ing for inter­net ser­vice, which costs $110 a month for Starlink’s stan­dard plan and $500 a month for its pre­mi­um plan avail­able with the more expen­sive ter­mi­nals.

The Wash­ing­ton Post also reports that USAID agreed to pay more than $800,000 for trans­porta­tion of the 5,000 ter­mi­nals that were sent to Ukraine through what the agency now calls a “pub­lic-pri­vate part­ner­ship.”

...

It does appear like SpaceX did make a sig­nif­i­cant char­i­ta­ble dona­tion to Ukraine. USAID told the Post in a state­ment that the “deliv­ery of Star­link ter­mi­nals were made pos­si­ble by a range of stake­hold­ers, whose com­bined con­tri­bu­tions val­ued over $15 mil­lion,” and its orig­i­nal press release said that SpaceX donat­ed $10 mil­lion worth of equip­ment in ser­vice. But, as with many oth­er SpaceX projects, the com­pa­ny does appear to have got­ten a sig­nif­i­cant amount of pub­lic fund­ing for the project.

———-

“The US gov­ern­ment report­ed­ly paid to send Star­link ter­mi­nals to Ukraine” by Mitchell Clark; The Verge; 04/08/2022

“Despite SpaceX imply­ing that the US didn’t give mon­ey to send Star­link ter­mi­nals to Ukraine in March, a report from The Wash­ing­ton Post reveals that the gov­ern­ment actu­al­ly paid mil­lions of dol­lars for equip­ment and trans­porta­tion. The report found that the US Agency for Inter­na­tion­al Devel­op­ment, or USAID, paid $1,500 apiece for 1,333 ter­mi­nals, adding up to around $2 mil­lion. USAID dis­closed the num­ber of ter­mi­nals it bought from the com­pa­ny in a press release from ear­ly April that has since been altered to remove men­tions of the pur­chase.”

For what­ev­er rea­son, USAID decid­ed to alter its press release on the ‘pub­lic-pri­vate’ part­ner­ship it start­ed with SpaceX to deliv­er 5,000 Star­link ter­mi­nals to Ukraine. Why the alter­ation? It’s a mys­tery, along with the mys­tery of whether or not the $1,500 USAID was pay­ing for these units was for the $600 ter­mi­nals or the more advanced $2,500 ter­mi­nals. If it was $1,500 going towards $2,500 ter­mi­nals, well, ok, that well be an obvi­ous sub­sidy towards SpaceX’s ‘char­i­ta­ble con­tri­bu­tions’. But if it was $1,500 going towards the $600 units, you have to won­der what exact­ly was going on here:

...
Accord­ing to space reporter Joey Roulette, SpaceX donat­ed 3,667 ter­mi­nals to Ukraine, or around $10 mil­lion worth, after also fac­tor­ing in the three months of data it pro­vid­ed with the ter­mi­nals. How­ev­er, it’s pos­si­ble this dona­tion was par­tial­ly sub­si­dized by USAID’s $1,500-a-unit pur­chase. The Wash­ing­ton Post says it’s unclear whether Ukraine received stan­dard ter­mi­nals, which SpaceX cur­rent­ly charges $600 for, or the advanced ter­mi­nals, which were announced in Feb­ru­ary and cost $2,500.

Looks like USAID edit­ed their press release and delet­ed the val­ue of SpaceX’s dona­tion, as well as the num­ber of ter­mi­nals the agency said SpaceX pro­vid­ed on its own dime. Here’s the pre­vi­ous ver­sion vs. cur­rent ver­sion: https://t.co/ws1urZu4VM pic.twitter.com/1eZxfQ6DPA— Joey Roulette (@joroulette) April 6, 2022

If USAID pur­chased reg­u­lar ter­mi­nals, it paid $900 over retail cost per unit (adding up to around $1.2 mil­lion) — notably, a recent report said that the ter­mi­nals cost SpaceX around $1,000 to build. It is pos­si­ble that USAID was also pay­ing for inter­net ser­vice, which costs $110 a month for Starlink’s stan­dard plan and $500 a month for its pre­mi­um plan avail­able with the more expen­sive ter­mi­nals.
...

And it’s the mys­tery of that rela­tion­ship between SpaceX and USAID that brings us to the fol­low arti­cle from back in August of last year about a pro­found­ly dis­turb­ing study of the impact SpaceX is already hav­ing on the sta­tus of space junk and orbital close encoun­ters. As researchers found, rough­ly half of the ~1,600 close encoun­ters mea­sured week­ly involved Star­link satel­lites. Half. We’re talk­ing about a satel­lite con­stel­la­tion that did­n’t event exist sev­er­al years ago. It now accounts for half of the orbital close encoun­ters. And as the arti­cle notes, only around 1,700 Star­link satel­lites had been launched by that point last year. The ulti­mate plan is the cre­ation of a orbital net­work that con­sists of tens of thou­sands of these Star­link microsatel­lites. That’s why these researchers are pre­dict­ing that Star­link satel­lites are on track to account for 90% of orbital close encoun­ters in com­ing years.

So Elon Musk is appar­ent­ly devel­op­ing a monop­oly on orbital close encoun­ters. And no one appears to be doing any­thing to stop it. Quite the con­trary, Star­link is involved with ‘pub­lic-pri­vate part­ner­ships’ with the US gov­ern­ment. And that’s why the sto­ries about Star­link’s USAID-spon­sored role in the war in Ukraine and the grow­ing threat it pos­es to the plan­e­tary orbital space are real­ly part of the same sto­ry:

Space.com

SpaceX Star­link satel­lites respon­si­ble for over half of close encoun­ters in orbit, sci­en­tist says

By Tereza Pul­taro­va
pub­lished August 18, 2021

Star­link satel­lites might soon be involved in 90% of close encoun­ters between two space­craft in low Earth orbit.

Oper­a­tors of satel­lite con­stel­la­tions are con­stant­ly forced to move their satel­lites because of encoun­ters with oth­er space­craft and pieces of space junk. And, thanks to SpaceX’s Star­link satel­lites, the num­ber of such dan­ger­ous approach­es will con­tin­ue to grow, accord­ing to esti­mates based on avail­able data.

SpaceX’s Star­link satel­lites alone are involved in about 1,600 close encoun­ters between two space­craft every week, that’s about 50 % of all such inci­dents, accord­ing to Hugh Lewis, the head of the Astro­nau­tics Research Group at the Uni­ver­si­ty of Southamp­ton, U.K. These encoun­ters include sit­u­a­tions when two space­craft pass with­in a dis­tance of 0.6 miles (1 kilo­me­ter) from each oth­er.

Lewis, Europe’s lead­ing expert on space debris, makes reg­u­lar esti­mates of the sit­u­a­tion in orbit based on data from the Socrates (Satel­lite Orbital Con­junc­tion Reports Assess­ing Threat­en­ing Encoun­ters in Space ) data­base. This tool, man­aged by Celestrack, pro­vides infor­ma­tion about satel­lite orbits and mod­els their tra­jec­to­ries into the future to assess col­li­sion risk.

Lewis pub­lish­es reg­u­lar updates on Twit­ter and has seen a wor­ry­ing trend in the data that reflects the fast deploy­ment of the Star­link con­stel­la­tion.

“I have looked at the data going back to May 2019 when Star­link was first launched to under­stand the bur­den of these mega­con­stel­la­tions,” Lewis told Space.com. “Since then, the num­ber of encoun­ters picked up by the Socrates data­base has more than dou­bled and now we are in a sit­u­a­tion where Star­link accounts for half of all encoun­ters.”

The cur­rent 1,600 close pass­es include those between two Star­link satel­lites. Exclud­ing these encoun­ters, Star­link satel­lites approach oth­er oper­a­tors’ space­craft 500 times every week.

[see graph]

In com­par­i­son, Star­link’s com­peti­tor OneWeb, cur­rent­ly fly­ing over 250 satel­lites, is involved in 80 close pass­es with oth­er oper­a­tors’ satel­lites every week, accord­ing to Lewis’ data.

And the sit­u­a­tion is bound to get worse. Only 1,700 satel­lites of an expect­ed con­stel­la­tion of tens of thou­sands have been placed into orbit so far. Once SpaceX launch­es all 12,000 satel­lites of its first gen­er­a­tion con­stel­la­tion, Star­link satel­lites will be involved in 90% of all close approach­es, Lewis’ cal­cu­la­tions sug­gest.

[see graph]

The risk of col­li­sion

Siemak Hesar, CEO and co-founder of Boul­der, Col­orado, based Kay­han Space, con­firms the trend. His com­pa­ny, which devel­ops a com­mer­cial autonomous space traf­fic man­age­ment sys­tem, esti­mates that on aver­age, an oper­a­tor man­ag­ing about 50 satel­lites will receive up to 300 offi­cial con­junc­tion alerts a week. These alerts include encoun­ters with oth­er satel­lites as well as pieces of debris. Out of these 300 alerts, up to ten might require oper­a­tors to per­form avoid­ance maneu­vers, Hesar told Space.com.

Kay­han Space bases their esti­mates on data pro­vid­ed by the U.S. Space Sur­veil­lance Net­work. This net­work of radars and tele­scopes, man­aged by the U.S. Space Force, close­ly mon­i­tors about 30,000 live and defunct satel­lites and pieces of debris down to the size of 4 inch­es (10 cen­time­ters) and pro­vides the most accu­rate loca­tion data of the orbit­ing objects.

The size of this cat­a­log is expect­ed to increase ten times in the near future, Hesar added, part­ly due to the growth of mega­con­stel­la­tions, such as Star­link, and part­ly as sen­sors improve and enable detec­tion of even small­er objects. The more objects in the cat­a­log mean more dan­ger­ous­ly close encoun­ters.

“This prob­lem is real­ly get­ting out of con­trol,” Hesar said. “The process­es that are cur­rent­ly in place are very man­u­al, not scal­able, and there is not enough infor­ma­tion shar­ing between par­ties that might be affect­ed if a col­li­sion hap­pens.”

...

Bad deci­sions

Despite the con­cerns, only three con­firmed orbital col­li­sions have hap­pened so far. Ear­li­er this week, astro­physi­cist and satel­lite track­er Jonathan McDow­ell, who’s based at the Har­vard-Smith­son­ian Cen­ter for Astro­physics in Cam­bridge, Mass­a­chu­setts, found evi­dence in Space-Track data that the Chi­nese mete­o­ro­log­i­cal satel­lite Yun­hai 1–02, which dis­in­te­grat­ed in March this year, was actu­al­ly hit by a piece of space debris.

The worst known space col­li­sion in his­to­ry took place in Feb­ru­ary 2009 when the U.S. telecom­mu­ni­ca­tion satel­lite Irid­i­um 33 and Rus­si­a’s defunct mil­i­tary satel­lite Kos­mos-2251 crashed at the alti­tude of 490 miles (789 kilo­me­tres). The inci­dent spawned over 1,000 pieces of debris larg­er than 4 inch­es (10 cm). Many of these frag­ments were then involved in fur­ther orbital inci­dents.

Lewis is con­cerned that with the num­ber of close pass­es grow­ing, the risk of oper­a­tors at some point mak­ing a wrong deci­sion will grow as well. Avoid­ance maneu­vers cost fuel, time and effort. Oper­a­tors, there­fore, always care­ful­ly eval­u­ate such risks. A deci­sion not to make an avoid­ance maneu­ver fol­low­ing an alert, such as that made by Irid­i­um in 2009, could, how­ev­er, clut­ter the orbital envi­ron­ment for years and decades.

“In a sit­u­a­tion when you are receiv­ing alerts on a dai­ly basis, you can’t maneu­ver for every­thing,” Lewis said. “The maneu­vers use pro­pel­lant, the satel­lite can­not pro­vide ser­vice. So there must be some thresh­old. But that means you are accept­ing a cer­tain amount of risk. The prob­lem is that at some point, you are like­ly to make a wrong deci­sion.”

Hesar said that uncer­tain­ties in the posi­tions of satel­lites and pieces of debris are still con­sid­er­able. In case of oper­a­tional satel­lites, the error could be up to 330 feet (100 meters) large. When it comes to a piece of debris, the uncer­tain­ty about its exact posi­tion might be in the order of a mile or more.

“This object can be any­where in this bub­ble of mul­ti­ple kilo­me­tres,” Hesar said. “At this point, and for the fore­see­able future, avoid­ance is our best recourse. Peo­ple that say ‘I’m going to take the risk’, in my hum­ble opin­ion, that’s an irre­spon­si­ble thing to do.”

Star­link monop­oly

Lewis is con­cerned about the grow­ing influ­ence of a sin­gle actor — Star­link — on the safe­ty of orbital oper­a­tions. Espe­cial­ly, he says, as the space­flight com­pa­ny has entered the satel­lite oper­a­tions world only recent­ly.

“We place trust in a sin­gle com­pa­ny, to do the right thing,” Lewis said. “We are in a sit­u­a­tion where most of the maneu­vers we see will involve Star­link. They were a launch provider before, now they are the world’s biggest satel­lite oper­a­tor, but they have only been doing that for two years so there is a cer­tain amount of inex­pe­ri­ence.”

SpaceX relies on an autonomous col­li­sion avoid­ance sys­tem to keep its fleet away from oth­er space­craft. That, how­ev­er, could some­times intro­duce fur­ther prob­lems. The auto­mat­ic orbital adjust­ments change the fore­cast­ed tra­jec­to­ry and there­fore make col­li­sion pre­dic­tions more com­pli­cat­ed, accord­ing to Lewis.

“Star­link does­n’t pub­li­cize all the maneu­vers that they’re mak­ing, but it is believed that they are mak­ing a lot of small cor­rec­tions and adjust­ments all the time,” Lewis said. “But that caus­es prob­lems for every­body else because no one knows where the satel­lite is going to be and what it is going to do in the next few days.”

———-

“SpaceX Star­link satel­lites respon­si­ble for over half of close encoun­ters in orbit, sci­en­tist says” by Tereza Pul­taro­va; Space.com; 08/18/2021

“SpaceX’s Star­link satel­lites alone are involved in about 1,600 close encoun­ters between two space­craft every week, that’s about 50 % of all such inci­dents, accord­ing to Hugh Lewis, the head of the Astro­nau­tics Research Group at the Uni­ver­si­ty of Southamp­ton, U.K. These encoun­ters include sit­u­a­tions when two space­craft pass with­in a dis­tance of 0.6 miles (1 kilo­me­ter) from each oth­er.”

Half of cur­rent space close encoun­ters today involve the Star­link con­stel­la­tion of satel­lites, some­thing that did­n’t exist a few years ago. In oth­er words, we don’t need to sim­ply wor­ry about these micro satel­lites cause a col­li­sion and gen­er­at­ing space junk. These things already are space junk. And 1,700 of these things have been launched so far. The plan is to put tens of thou­sands of these micro satel­lites into orbit. So we’re just expe­ri­enc­ing a taste of the orbital traf­fic jams yet to come.

And it’s not just like these close encoun­ters just involve Star­link satel­lites threat­en­ing a non-Star­link satel­lite. Some of these close encoun­ters involve two Star­link satel­lites. The Star­link con­stel­la­tion is lit­er­al­ly a threat to itself and it’s not even close to be ful­ly launched yet. That’s why these experts are pre­dict­ing that 90% of the close encoun­ters in the future are going to involve Star­link satel­lites. It’s a space junk monop­oly, seem­ing­ly being built with the endorse­ment of the US gov­ern­ment:

...
“I have looked at the data going back to May 2019 when Star­link was first launched to under­stand the bur­den of these mega­con­stel­la­tions,” Lewis told Space.com. “Since then, the num­ber of encoun­ters picked up by the Socrates data­base has more than dou­bled and now we are in a sit­u­a­tion where Star­link accounts for half of all encoun­ters.”

The cur­rent 1,600 close pass­es include those between two Star­link satel­lites. Exclud­ing these encoun­ters, Star­link satel­lites approach oth­er oper­a­tors’ space­craft 500 times every week.

...

And the sit­u­a­tion is bound to get worse. Only 1,700 satel­lites of an expect­ed con­stel­la­tion of tens of thou­sands have been placed into orbit so far. Once SpaceX launch­es all 12,000 satel­lites of its first gen­er­a­tion con­stel­la­tion, Star­link satel­lites will be involved in 90% of all close approach­es, Lewis’ cal­cu­la­tions sug­gest.
...

But as these experts point out, the grow­ing threat posed by the Star­link con­stel­la­tion isn’t just the direct threat of a space col­li­sion. There’s also the threat that this abun­dance of close encoun­ters is going to cause satel­lite oper­a­tors to because far more risk tol­er­ant than they should. Repo­si­tion­ing satel­lites takes time and fuel. Satel­lite oper­ates are going to be forced to make judge­ment calls on whether or not a close encounter warn­ing is worth respond­ing to and it’s just a mat­ter of time before they make a mis­take. The kind of mis­take that can have cas­cad­ing costs:

...
Kay­han Space bases their esti­mates on data pro­vid­ed by the U.S. Space Sur­veil­lance Net­work. This net­work of radars and tele­scopes, man­aged by the U.S. Space Force, close­ly mon­i­tors about 30,000 live and defunct satel­lites and pieces of debris down to the size of 4 inch­es (10 cen­time­ters) and pro­vides the most accu­rate loca­tion data of the orbit­ing objects.

The size of this cat­a­log is expect­ed to increase ten times in the near future, Hesar added, part­ly due to the growth of mega­con­stel­la­tions, such as Star­link, and part­ly as sen­sors improve and enable detec­tion of even small­er objects. The more objects in the cat­a­log mean more dan­ger­ous­ly close encoun­ters.

“This prob­lem is real­ly get­ting out of con­trol,” Hesar said. “The process­es that are cur­rent­ly in place are very man­u­al, not scal­able, and there is not enough infor­ma­tion shar­ing between par­ties that might be affect­ed if a col­li­sion hap­pens.”

...

Lewis is con­cerned that with the num­ber of close pass­es grow­ing, the risk of oper­a­tors at some point mak­ing a wrong deci­sion will grow as well. Avoid­ance maneu­vers cost fuel, time and effort. Oper­a­tors, there­fore, always care­ful­ly eval­u­ate such risks. A deci­sion not to make an avoid­ance maneu­ver fol­low­ing an alert, such as that made by Irid­i­um in 2009, could, how­ev­er, clut­ter the orbital envi­ron­ment for years and decades.

“In a sit­u­a­tion when you are receiv­ing alerts on a dai­ly basis, you can’t maneu­ver for every­thing,” Lewis said. “The maneu­vers use pro­pel­lant, the satel­lite can­not pro­vide ser­vice. So there must be some thresh­old. But that means you are accept­ing a cer­tain amount of risk. The prob­lem is that at some point, you are like­ly to make a wrong deci­sion.”

...

SpaceX relies on an autonomous col­li­sion avoid­ance sys­tem to keep its fleet away from oth­er space­craft. That, how­ev­er, could some­times intro­duce fur­ther prob­lems. The auto­mat­ic orbital adjust­ments change the fore­cast­ed tra­jec­to­ry and there­fore make col­li­sion pre­dic­tions more com­pli­cat­ed, accord­ing to Lewis.

“Star­link does­n’t pub­li­cize all the maneu­vers that they’re mak­ing, but it is believed that they are mak­ing a lot of small cor­rec­tions and adjust­ments all the time,” Lewis said. “But that caus­es prob­lems for every­body else because no one knows where the satel­lite is going to be and what it is going to do in the next few days.”
...

How long before the world’s satel­lite oper­a­tors hit ‘close encounter fatigue’ and just stop mov­ing their satel­lites out of the way? The only gov­ern­ment in a posi­tion to pre­vent that even­tu­al­i­ty is sub­si­diz­ing it instead, so we’ll find out even­tu­al­ly.

Fol­low­ing up on the role Elon Musk’s Star­link is play­ing in the con­flict in Ukraine — sub­si­dized by USAID — and the poten­tial risks of a cas­cad­ing orbital cat­a­stro­phe (Kessler’s Syn­drome) that comes with the mil­i­ta­riza­tion of the Star­link low orbit con­stel­la­tion of mini-satel­lites, here’s a pair of arti­cles that should serve as a warn­ing that we should prob­a­bly expect the Star­link sys­tem to be treat­ed as a mil­i­tary tar­get one of these days, with all of the cas­cad­ing con­se­quences that could arise from that. Because as the arti­cles describe, Star­link has already come under a kind of Russ­ian attack. Specif­i­cal­ly, a sig­nal jam­ming effort that appears to have worked for at least a few hours in Ukraine before Star­link was able to issue a patch that fixed the prob­lem.

The attack took place back in ear­ly March. We aren’t giv­en any details on the Russ­ian sig­nal jam­ming attack, but it was pre­sum­ably some sort of elec­tron­ic war­fare mea­sure that dis­rupt­ed the the abil­i­ty of the Star­link ter­mi­nals locat­ed on the ground in Ukraine to com­mu­ni­cate with the satel­lites. We can also infer that the fix did­n’t require any updates to the ter­mi­nals them­selves since they would­n’t have been able to receive the updates. So some sort of update was deliv­ered to the soft­ware oper­at­ing the satel­lites them­selves that fixed the jam­ming. That’s about all we know about the inci­dent.

Over­all, it sounds like a rel­a­tive­ly sim­ple form of elec­tron­ic war­fare. It does­n’t sound like the attack actu­al­ly hack­ing the the soft­ware oper­at­ing these satel­lites. But the fact that the coun­ter­mea­sure for the attack involved a rapid soft­ware patch under­scores the basic fact that this con­stel­la­tion of satel­lites has the abil­i­ty to have its soft­ware rapid­ly remote­ly updat­ed. Because of course it has the capa­bil­i­ty. It’s an absolute neces­si­ty for man­ag­ing a grow­ing chaot­ic clus­ter. Don’t for­get what researchers con­clud­ed last year: Star­link satel­lites are cur­rent­ly respon­si­ble for rough­ly half the close encounter events and will like­ly be the source of 90% of close encoun­ters by the time SpaceX is done launch­ing the tens of thou­sands of mini-satel­lites it has planned. Some of the close encoun­ters involve two Star­link satel­lites careen­ing towards each oth­er. Hav­ing the abil­i­ty to remote­ly update the Star­link soft­ware and remote­ly adjust the orbits of each one of those satel­lites is an absolute neces­si­ty.

But that neces­si­ty for remote­ly pilot­ing this unprece­dent­ed satel­lite clus­ter also obvi­ous­ly pos­es a hack­ing risk. Yes, there’s no indi­ca­tion that and Star­link satel­lites were hacked as part of this sig­nal jam­ming cam­paign. But the poten­tial is obvi­ous­ly there. It’s not like satel­lites are immune to hack­ing. Quite the con­trary. Satel­lites are noto­ri­ous­ly easy to hack.

And not only are there plen­ty of exam­ples of hack­ers hack­ing satel­lites for fun, don’t for­get that you don’t nec­es­sar­i­ly need to hack the satel­lite direct­ly. Hack­ing the satel­lite oper­a­tor could poten­tial­ly give you remote access to those satel­lites too. Rus­si­a’s mil­i­tary was accused of hack­ing Ukrain­ian satel­lite com­pa­ny Viasat at the begin­ning of the con­flict. We don’t have any indi­ca­tion that the hack gave Rus­sia con­trol over Viasat’s satel­lites. But as we’ve seen with the Solar­Winds hacks, once a sophis­ti­cat­ed hack­er is allow into a cor­po­rate net­work it can be very dif­fi­cult to get them out. Was Star­link hit by the Solar­Winds hack? How about some Star­link con­trac­tors? It only takes one com­pro­mised part­ner.

Final­ly, also recall how Star­link relies in part on auto­mat­ed orbital adjust­ments to avoid col­li­sion. Imag­ine a hack that sends faulty code han­dling that part of Star­link’s func­tion­al­i­ty. You could the­o­ret­i­cal­ly send the entire clus­ter careen­ing into itself and the rest of the satel­lites in low earth orbit.

And that’s all why the suc­cess­ful repelling of a Russ­ian sig­nal jam­ming attempt should­n’t nec­es­sar­i­ly be a relief for any­one con­cerned about the poten­tial risk these con­stel­la­tions of microsatel­lites pose to human­i­ty’s abil­i­ty to oper­ate in space. Yes, this par­tic­u­lar attack did­n’t suc­ceed. But with Star­link, we’re still one suc­cess­ful hack away from an orbital cat­a­stro­phe:

C4ISR.net

SpaceX shut down a Russ­ian elec­tro­mag­net­ic war­fare attack in Ukraine last month — and the Pen­ta­gon is tak­ing notes

By Stephen Losey
Wednes­day, Apr 20, 2022

WASHINGTON — Russia’s halt­ing efforts to con­duct elec­tro­mag­net­ic war­fare in Ukraine show how impor­tant it is to quick­ly respond, and imme­di­ate­ly shut down, such attacks, Pen­ta­gon experts said Wednes­day.

But the U.S. needs to get much bet­ter at its own EW rapid response, they said dur­ing the C4ISRNET Con­fer­ence Wednes­day — and can learn a lot from how the pri­vate sec­tor has han­dled these sit­u­a­tions.

Brig. Gen. Tad Clark, direc­tor of the Air Force’s elec­tro­mag­net­ic spec­trum supe­ri­or­i­ty direc­torate, said mod­ern wars will increas­ing­ly involve elec­tro­mag­net­ic war­fare, par­tic­u­lar­ly to shape the bat­tle­field when con­flicts begin.

Dave Trem­per, direc­tor of elec­tron­ic war­fare for the Office of the Sec­re­tary of Defense, point­ed to SpaceX’s abil­i­ty last month to swift­ly stymie a Russ­ian effort to jam its Star­link satel­lite broad­band ser­vice, which was keep­ing Ukraine con­nect­ed to the Inter­net. SpaceX founder Elon Musk steered thou­sands of Star­link ter­mi­nals to Ukraine after an offi­cial sent him a tweet ask­ing for help keep­ing the besieged coun­try online.

“The next day [after reports about the Russ­ian jam­ming effort hit the media], Star­link had slung a line of code and fixed it,” Trem­per said. “And sud­den­ly that [Russ­ian jam­ming attack] was not effec­tive any­more. From [the] EW technologist’s per­spec­tive, that is fan­tas­tic … and how they did that was eye-water­ing to me.”

The gov­ern­ment, on the oth­er hand, has a “sig­nif­i­cant time­line to make those types of cor­rec­tions” as it mud­dles through analy­ses of what hap­pened, decides how to fix it and gets a con­tract in place for the fix.

“We need to be able to have that agili­ty,” Trem­per said. “We need to be able to change our elec­tro­mag­net­ic pos­ture to be able to change, very dynam­i­cal­ly, what we’re try­ing to do with­out los­ing capa­bil­i­ty along the way.”

Redun­dan­cy is also crit­i­cal so the U.S. could keep oper­at­ing on anoth­er sys­tem if an EW attack suc­ceed­ed at knock­ing one out, Trem­per said.

The U.S. needs to think a lot more inno­v­a­tive­ly when it comes to build­ing new EW equip­ment, Clark said. It won’t be enough to just buy upgrad­ed ver­sions of lega­cy sys­tems, he said — the U.S. has to come up with new sys­tems that allow for much greater resilience and speed.

This includes incor­po­rat­ing arti­fi­cial intel­li­gence and machine learn­ing into next-gen­er­a­tion sys­tems to be able to respond faster, he said. Increased use of dig­i­tal engi­neer­ing can also help the mil­i­tary mod­el new equip­ment with a com­put­er and work out the kinks before going through the time-con­sum­ing typ­i­cal acqui­si­tion and test­ing process.

...

Trem­per said the Pen­ta­gon expect­ed a “much stronger” EW show­ing from Rus­sia — but cau­tioned that isn’t to say all of Russia’s efforts have failed.

————

“SpaceX shut down a Russ­ian elec­tro­mag­net­ic war­fare attack in Ukraine last month — and the Pen­ta­gon is tak­ing notes” by Stephen Losey; C4ISR.net; 04/20/2022

““The next day [after reports about the Russ­ian jam­ming effort hit the media], Star­link had slung a line of code and fixed it,” Trem­per said. “And sud­den­ly that [Russ­ian jam­ming attack] was not effec­tive any­more. From [the] EW technologist’s per­spec­tive, that is fan­tas­tic … and how they did that was eye-water­ing to me.””

A soft­ware update end­ed the attack. On the one hand, that’s a nice sign for Star­link’s robust­ness in the face of an out­side attack like a jam­ming sig­nal. But it’s also a reminder that if hack­ers in the future man­age to hack Star­link’s own sys­tems they just might find them­selves with the capac­i­ty to update Star­link’s satel­lite soft­ware. So when Elon Musk tweet­ed out that “SpaceX repri­or­i­tized to cyber defense & over­com­ing sig­nal jam­ming”, in response to the inci­dent back in March, let’s hope that includes pro­tect­ing not just the satel­lites but all of the sys­tems task with remote­ly con­trol­ling these satel­lites:

Space.com

Elon Musk says SpaceX focus­ing on cyber defense after Star­link sig­nals jammed near Ukraine con­flict areas

Star­ship and Star­link V2 progress will be delayed, Musk said.

By Tariq Malik
pub­lished March 05, 2022

SpaceX founder and CEO Elon Musk said Fri­day that his com­pa­ny is now focus­ing on cyber defense and over­com­ing sig­nal jam­ming of its Star­link inter­net satel­lites amid Rus­si­a’s ongo­ing inva­sion of Ukraine.

Musk and SpaceX sent Star­link ter­mi­nals to Ukraine at the request of a gov­ern­ment offi­cial after inter­net ser­vice was dis­rupt­ed across the coun­try by the Russ­ian inva­sion. A ship­ment of Star­link ground ter­mi­nals, which use an anten­na and ter­mi­nal to access the satel­lite broad­band ser­vice, arrived in Ukraine by Mon­day Feb. 28). With the ter­mi­nals in use, SpaceX is work­ing to keep them online, Musk said.

“Some Star­link ter­mi­nals near con­flict areas were being jammed for sev­er­al hours at a time,” Musk wrote in a Twit­ter state­ment Fri­day (March 1). “Our lat­est soft­ware update bypass­es the jam­ming.”

SpaceX repri­or­i­tized to cyber defense & over­com­ing sig­nal jam­ming. Will cause slight delays in Star­ship & Star­link V2.— Elon Musk (@elonmusk) March 5, 2022

Musk lat­er said SpaceX is shift­ing its focus to keep­ing its Star­link ser­vice unin­ter­rupt­ed in Ukraine and like­ly else­where.

“SpaceX repri­or­i­tized to cyber defense & over­com­ing sig­nal jam­ming,” he wrote Fri­day. Musk quipped that the mea­sures were a bit of unex­pect­ed qual­i­ty assur­ance work for the Star­link sys­tem.

Musk also said the Star­link work “will cause slight delays in Star­ship & Star­link V2.”

SpaceX’s Star­ship is a giant reusable space­craft designed to use a huge reusable boost­er called Super Heavy to launch mis­sions into deep space. NASA has tapped the Star­ship vehi­cle to land astro­nauts on the moon for its Artemis pro­gram. SpaceX is hop­ing the launch the first orbital flight of an uncrewed Star­ship in the next few months. Star­link V2 is SpaceX’s next-gen­er­a­tion Star­link sys­tem that includes laser links between satel­lites and oth­er enhance­ments.

Impor­tant warn­ing: Star­link is the only non-Russ­ian com­mu­ni­ca­tions sys­tem still work­ing in some parts of Ukraine, so prob­a­bil­i­ty of being tar­get­ed is high. Please use with cau­tion.— Elon Musk (@elonmusk) March 3, 2022

After deliv­er­ing Star­link ter­mi­nals to Ukraine, Musk cau­tioned that the sys­tem could make its users vul­ner­a­ble to Russ­ian mil­i­tary attacks.

...

On Thurs­day (March 3), SpaceX sent its lat­est batch of Star­link satel­lites into orbit on a Fal­con 9 rock­et. That mis­sion launched 47 new Star­link satel­lites in orbit from Pad 39A at NASA’s Kennedy Space Cen­ter. To date, SpaceX has launched more than 2,000 satel­lites into orbit, with plans for an ini­tial mega­con­stel­la­tion of 12,000 to pro­vide glob­al broad­band cov­er­age.

———

“Elon Musk says SpaceX focus­ing on cyber defense after Star­link sig­nals jammed near Ukraine con­flict areas” by Tariq Malik; Space.com; 03/05/2022

““SpaceX repri­or­i­tized to cyber defense & over­com­ing sig­nal jam­ming,” he wrote Fri­day. Musk quipped that the mea­sures were a bit of unex­pect­ed qual­i­ty assur­ance work for the Star­link sys­tem.”

SpaceX had to repri­or­i­tize not just over­com­ing the direct attack of sig­nal jam­ming, but also cyber defense. It’s an implic­it acknowl­edge­ment that the Star­link sys­tem’s vul­ner­a­bil­i­ties don’t just involve some sort of direct phys­i­cal attack. Star­link can poten­tial­ly get hacked too, whether we’re talk­ing about the direct hack­ing of these satel­lites or the indi­rect hack­ing of the Star­link com­mand and con­trol cen­ters where these kind of remote soft­ware updates can get pushed to the net­work.

So with Star­link hav­ing already been weaponized for bat­tle­field uses and already hav­ing come under at least an indi­rect dis­rup­tion of its ser­vices in response to that weaponiza­tion, we have to ask: how high was cyber defense on the pri­or­i­ty list when SpaceX was orig­i­nal design­ing the Star­link sys­tem? Don’t for­get that Star­link is a plat­form that’s already been rushed through with­out a num­ber of oth­er prop­er safe­ty assess­ments, like the basic assess­ment of whether or not it’s safe to sud­den­ly launch thou­sands of microsatel­lites into low earth orbit with­out trig­ger­ing some sort of Kessler Syn­drone cas­cade cat­a­stro­phe. Was cyber­se­cu­ri­ty also rushed through in the race to be the first com­pa­ny with a ‘mega­con­stel­la­tion’ of satel­lites in orbit? Star­link rep­re­sents a kind of orbital land grab, after all. How high a pri­or­i­ty was cyber­se­cu­ri­ty in this land grab? It’s a ques­tion that is quite lit­er­al­ly loom­ing over all of us. Well, loom­ing over most of us. If you hap­pen to be serv­ing a space sta­tion, the threat threat is more adja­cent.