Less than an hour before, Homeland Security Secretary Kristi Noem had posted a photo of the exact same scene, but in Noem’s version Levy Armstrong appeared composed, not crying in the least.

Seeking to determine if the White House version of the photo had been altered using artificial intelligence tools, we turned to Google’s SynthID — a detection mechanism that Google claims is able to discern whether an image or video was generated using Google’s own AI. We followed Google’s instructions and used its AI chatbot, Gemini, to see if the image contained SynthID forensic markers.

The results were clear: The White House image had been manipulated with Google’s AI. We published a story about it.

Related

White House Doctored Photo With AI to Make It Look Like an Activist Was Sobbing During Perp Walk

After posting the article, however, subsequent attempts to use Gemini to authenticate the image with SynthID produced different outcomes.

In our second test, Gemini concluded that the image of Levy Armstrong crying was actually authentic. (The White House doesn’t even dispute that the image was doctored. In response to questions about its X post, a spokesperson said, “The memes will continue.”)

In our third test, SynthID determined that the image was not made with Google’s AI, directly contradicting its first response.

At a time when AI-manipulated photos and videos are growing inescapable, these inconsistent responses raise serious questions about SynthID’s reliability to tell fact from fiction.

Initial SynthID Results

Google describes SynthID as a digital watermarking system. It embeds invisible markers into AI-generated images, audio, text or video created using Google’s tools, which it can then detect — proving whether a piece of online content is authentic.

“The watermarks are embedded across Google’s generative AI consumer products, and are imperceptible to humans — but can be detected by SynthID’s technology,” says a page on the site for DeepMind, Google’s AI division.

Google presents SynthID as having what in the realm of digital watermarking is known as “robustness” — it claims to be able to detect the watermarks even if an image undergoes modifications, such as cropping or compression. Therefore, an image manipulated with Google’s AI should contain detectable watermarks even if it has been saved multiple times or posted on social media.

Google steers those who want to use SynthID toward its Gemini AI chatbot, which they can prompt with questions about the authenticity of digital content.

“Want to check if an image or video was generated, or edited, by Google AI? Ask Gemini,” the SynthID landing page says.

We decided to do just that.

We saved the image file that the official White House account posted on X, bearing the filename G_R3H10WcAATYht.jfif, and uploaded it to Gemini. We asked whether SynthID detected the image had been generated with Google’s AI.

To test SynthID’s claims of robustness, we also uploaded a further cropped and re-encoded image, which we named imgtest2.jpg.

Finally, we uploaded a copy of the photo where Levy Armstrong was not crying, as previously posted by Noem. (In the above screenshot, Gemini refers to Noem’s photo as signal-2026-01-22-122805_002.jpeg because we downloaded it from the Signal messaging app).

“I’ve analyzed the images you provided,” wrote Gemini. “Based on the results from SynthID, all or part of the first two images were likely generated or modified with Google AI.”

“Technical markers within the files imgtest2.jpg and G_R3H10WcAATYht.jfif indicate the use of Google’s generative AI tools to alter the subject’s appearance,” the bot wrote. It also identified the version of the image posted by Noem as appearing to “be the original photograph.”

With confirmation from Google that its SynthID system had detected hidden forensic watermarks in the image, we reported in our story that the White House had posted an image that had been doctored with Google’s AI.

This wasn’t the only evidence the White House image wasn’t real; Levy Armstrong’s attorney told us that he was at the scene during the arrest and that she was not at all crying. The White House also openly described the image as a meme.

A Striking Reversal

A few hours after our story published, Google told us that they “don’t think we have an official comment to add.” A few minutes after that, a spokesperson for the company got back to us and said they could not replicate the result we got. They asked us for the exact files we uploaded. We provided them.

The Google spokesperson then asked, “Were you able to replicate it again just now?”

We ran the analysis again, asking Gemini to see if SynthID detected the image had been manipulated with AI. This time, Gemini failed to reference SynthID at all — despite the fact we followed Google’s instructions and explicitly asked the chatbot to use the detection tool by name. Gemini now claimed that the White House image was instead “an authentic photograph.”

It was a striking reversal considering Gemini previously said that the image contained technical markers indicating the use of Google’s generative AI. Gemini also said, “This version shows her looking stoic as she is being escorted by a federal agent” — despite our question addressing the version of the image depicting Levy Armstrong in tears.

Less than an hour later, we ran the analysis one more time, prompting Gemini to yet again use SynthID to check whether the image had been manipulated with Google’s AI. Unlike the second attempt, Gemini invoked SynthID as instructed. This time, however, it said, “Based on an analysis using SynthID, this image was not made with Google AI, though the tool cannot determine if other AI products were used.”

Google did not answer repeated questions about this discrepancy. In response to inquiries, the spokesperson continued to ask us to share the specific phrasing of the prompt that resulted in Gemini recognizing a SynthID marker in the White House image.

We didn’t store that language, but told Google it was a straightforward prompt asking Gemini to check whether SynthID detected the image as being generated with Google’s AI. We provided Google with information about our prompt and the files we used so the company could check its records of our queries in its Gemini and SynthID logs.

“We’re trying to understand the discrepancy,” said Katelin Jabbari, a manager of corporate communications at Google. Jabbari repeatedly asked if we could replicate the initial results, as “none of us here have been able to.”

After further back and forth following subsequent inquiries, Jabbari said, “Sorry, don’t have anything for you.”

Bullshit Detector?

Aside from Google’s proprietary tool, there is no easy way for users to test whether an image contains a SynthID watermark. That makes it difficult in this case to determine whether Google’s system initially detected the presence of a SynthID watermark in an image without one, or if subsequent tests missed a SynthID watermark in an image that actually contains one.

As AI become increasingly pervasive, the industry is trying to put behind its long history of being what researchers call a “bullshit generator.”

Supporters of the technology argue tools that can detect if something is AI will play a critical role establishing the common truth amid the pending flood of media generated or manipulated by AI. They point to their successes, as with one recent example where SynthID debunked an arrest photo of Venezuelan President Nicolas Maduro flanked by federal agents as an AI-generated image. The Google tool said the photo was bullshit.

If AI-detection technology fails to produce consistent responses, though, there’s reason to wonder who will call bullshit on the bullshit detector.

The post Google’s AI Detection Tool Can’t Decide if Its Own AI Made Doctored Photo of Crying Activist appeared first on The Intercept.

Google then purges these reviews and blocks the ability to write new ones, according to newly published research.

“Google actively suppresses carceral information, narratives, and digital trace data related to the U.S. Immigration and Customs Enforcement’s (ICE) detention facilities,” wrote Muira McCammon, a professor at Tulane University’s Department of Communication who has been scraping and analyzing self-reported accounts of ICE imprisonment posted to Google Maps since last year.

Much of this record is now permanently deleted, erasing detailed testimony of people ingested by the U.S. deportation machine.

Last August, McCammon aggregated 1,636 posts from 26 different ICE facilities where Google still allowed users to post reviews and discovered there were 91 other immigration jails on Google Maps with no reviews.

“Google had already turned off reviews for the remaining 91 facilities,” McCammon told The Intercept. “Almost all of the Google reviews that I originally collected in 2025 are now no longer available.”

Only four ICE facilities appear to still permit reviews on Google Maps, she said.

Pulling the reviews has come with no public notice or explanation.

In 2023, Google said in a corporate blog post that it reserved the right to delete Maps reviews for locations that are “consistently unhelpful, harmful, or off-topic. This includes places that people go to without choice or places only accessible to people stationed or assigned there—such as police stations and prisons.”

Unhelpful to whom, however, is unclear: While some of the reviews in question were statements of protest that could have been left by anyone, many appeared to come from people directly affected by ICE. They described a variety of issues that would be of great material interest to anyone detained in the facilities or for people close to them.

“Reviews help people make decisions on where to go,” Google spokesperson Genevieve Park said in a statement, “and we don’t show reviews on several categories of places where people go without choice,” citing the company’s policy of removing Google Maps reviews that are “unhelpful, harmful, or off-topic.”

In the past, Google has complied with requests to censor speech about ICE. In 2025, the company blocked Android users from downloading ICEBlock, an app that alerted people to crowdsourced sightings of ICE agents.

Related

Google Fulfilled ICE Subpoena Demanding Student Journalist’s Bank and Credit Card Numbers

Asked if the Department of Homeland Security, ICE’s parent agency, had communicated with Google about the reviews, Park said the takedowns were “unrelated—it was about enforcing our policy that prohibits reviews for places that people go to without choice.”

In a statement to The Intercept, a spokesperson for DHS denied that ICE requested any reviews of detention facilities be removed.

“Being in detention is a choice,” the Homeland Security spokesperson said, citing the department’s program to encourage an app-based system for self-deportation.

Many of the reviews McCammon aggregated were from despairing friends and relatives.

“My friend’s dad was sentenced to one month in jail,” read one review. “He served the time and then was moved to this facility for another 3 months. How is it not illegal to hold someone 3x longer than what they were sentenced to serve?”

A Spanish language review McCammon collected read, “You can barely hear the detainee when you make a phone call.”

Other posts reflect what apparent former detainees described as poor conditions, such as “The place is infested with roaches crawling all over the inmates body there,” or “Very bad experience i spent 2 months in this hell . . . 99% of staff is cruel and racist.”

One reviewer said many detainees weren’t able to communicate with ICE staff because of the language barrier:

I stayed there for more than a month. The food there was okay, but very unpalatable. They were all very kind. Due to the language barrier, there was a big barrier to communication. I felt that most people couldn’t express their thoughts and needs. Why didn’t they have a translator? ! Even if it’s not expensive. The air conditioner there was very cold, and many people, including me, caught cold from the cold. Even so, they were unwilling to turn up the air conditioner temperature. When I slept at night, I only had a thin blanket to cover myself, and it felt very cold.

Although Google’s stance appears to be that the censorship is based on a common-sense, consumer-centric view of Google Maps’ purpose, McCammon argues in her paper that these are essentially political decisions, made by “suppressing the circulation of carceral information and counter-narratives [and] calling into question how the company approaches the ideology of search, when it concerns the reputation of the U.S. administrative state.”

The post Google Is Censoring Reviews of ICE Detention Centers appeared first on The Intercept.

Musk’s break with OpenAI, which he co-founded in 2015, is in a sense a lawsuit about safety. He contends that Altman betrayed the company’s original nonprofit mission of safely and responsibly pursuing artificial intelligence for the public benefit by converting it into the revenue-maximizing behemoth it has become. According to Musk, the stakes of this are existential for the human race: “It could kill us all,” he testified on Tuesday. “We don’t want to have a ‘Terminator’ outcome.”

The AI safety community frequently invokes these dystopian scenarios to both warn the public about the technology’s risks and implicitly boast of its great power. While such a science-fiction future may lay ahead, these warnings overlook the deadly present. Artificial intelligence is already targeting humans with the blessing of Musk and his rivals.

Musk and others who caution about an uprising of sentient killer machines are anticipating the emergence of “artificial general intelligence,” an ill-defined form of superior machine reasoning that may never come to pass. But their fear that AI could kill us all is less hypothetical for those living in places targeted by the Trump administration’s global wars. In Iran, for instance, Anthropic’s Claude AI model “suggested hundreds of targets, issued precise location coordinates, and prioritized those targets according to importance,” according to the Washington Post.

“The risks of integrating frontier AI into the nation’s most lethal capabilities are already existential, both for civilians swept up in the violence and destruction of AI-enabled wars, and rank-and-file troops that have to live with the consequences of potentially unsafe weapons they can’t control,” Amoh Toh, senior counsel at Brennan Center’s Liberty and National Security Program, told The Intercept. “Existing AI models are already pushing policymakers and militaries toward nuclear escalation — there’s a real danger of Skynet-like outcomes even without a Skynet-style takeover.”

Related

OpenAI on Surveillance and Autonomous Killings: You’re Going to Have to Trust Us

Silicon Valley has widely embraced AI military contracts despite its worries over lethal AI. Amazon, OpenAI, Musk’s xAI, and Microsoft all earn money from selling large language model services to the Pentagon. Even Anthropic, accused of “betrayal” by War Secretary Pete Hegseth and declared a national supply chain risk for mounting the smallest of opposition to the Pentagon’s terms, is still keen to participate in the national kill chain. “Anthropic has much more in common with the Department of War than we have differences,” CEO Dario Amodei wrote in a blog post a week after the United States bombed an elementary school in Iran, killing more than 100 children.

Google offers a telling illustration of the industry’s increasing coziness with selling AI to the military. Following a 2018 employee revolt over Project Maven, a contract to help target Pentagon airstrikes, CEO Sundar Pichai pledged his company would swear off the business of killing. He wrote in a company blog post that Google would not pursue deals that could cause harm, including applications whose “principal purpose or implementation is to cause or directly facilitate injury to people.” He added: “These are not theoretical concepts, they are concrete standards that will actively govern our research and product development and will impact our business decisions.”

After watching AI help wage a war that has already killed over 1,700 Iranian civilians, Google this week sent a clear message: We want in. In a deal that makes explicit the extent to which company leadership has abandoned its AI principles, Google agreed to provide AI services to the Pentagon that allow for “classified workloads,” sensitive military work that encompasses tasks like intelligence analysis and targeting airstrikes, The Information reported.

According to the tech news outlet, the deal allows the U.S. military to use Google’s AI models for “any lawful government purpose” — a carveout that could allow any uses the administration deems legal. Take, for example, the Trump administration’s Operation Southern Spear, the ongoing aerial assassination program against civilian boats accused of drug trafficking that has killed more than 180 people to date. The campaign has been widely condemned as illegal under both international and U.S. law, but the administration has deemed its own actions legal through a Department of Justice memo that remains secret. On Friday, the Pentagon announced additional “lawful operational use” deals with Nvidia, Microsoft, and Amazon as well.

The Google contract reportedly includes a toothless and unenforceable provision gesturing at concerns over autonomous and spying. “We remain committed to the private and public sector consensus that AI should not be used for domestic mass surveillance or autonomous weaponry without appropriate human oversight,” the clause reportedly states.

“When I worked at Google, they would spend a lot of time punting into the future, promising a future that would never come,” said William Fitzgerald, a former Google employee who helped organize the 2018 worker-led campaign against the Maven contract. “‘Don’t regulate us or it’ll kill innovation.’ The talking point is the same today. The reality of Google’s work with the military is it’s part of a tech-military ecosystem that’s killing people today.”

Google spokesperson Kate Dreyer did not respond to questions about the contract’s language, instead touting how the company’s military work applies “to areas like logistics, cybersecurity, diplomatic translation, fleet maintenance, and the defense of critical infrastructure.”

There is little evidence the people in charge find this technology enticing because of its diplomatic translation prowess. In a January address to Musk’s employees at SpaceX, another Pentagon contractor, Hegseth explained how “an embrace of AI” would make the military “more lethal.”

Related

AI’s Imperial Agenda

Musk and Altman, though foes at the moment, can at least find common ground in their support of Hegseth. Musk, a longtime defense contractor, similarly wraps himself in the flag, tweeting in 2023, “I will fight for and die in America.” Altman, who once expressed skepticism toward military work, now frames OpenAI’s mission in terms of patriotic nationalism. (In 2024, The Intercept sued OpenAI in federal court over the company’s use of copyrighted articles to train its chatbot ChatGPT. The case is ongoing.)

Between Musk’s courtroom visions of the apocalypse and Google’s plunge into classified workloads, the week’s news illustrates the disjointed state of AI industry ethics, where executives say they’re terrified of the technology killing by accident, while wholly supportive of using it to kill on purpose.

Though AI executives clearly find this a virtuous revenue stream, some of the people who actually built the technology do not. Andreas Kirsch, a research scientist at Google’s pioneering DeepMind laboratory that produced much of the work on which xAI and Anthropic rely, responded to this week’s news with dismay: “I’m speechless at Google signing a deal to use our AI models for classified tasks. Frankly, it is shameful,” he wrote on X. Alex Turner, a DeepMind colleague of Kirsch’s, described the contract in a single word: “Shameful.”

The post Musk Warns of Killer AI — While He and the Rest of Silicon Valley Cash In on AI That Kills appeared first on The Intercept.

A Google digital watermarking system initially flagged that the image had been manipulated with Google’s AI tools. Later tests after the publication of this article resulted in different responses — raising questions about the consistency of Google’s SynthID system.

In the original photo, local civil rights activist Nekima Levy Armstrong was shown being escorted by authorities after her arrest in connection to a protest against U.S. Immigration and Customs Enforcement at Cities Church in Saint Paul, Minnesota. 

The version published by the White House’s official X account showed an image that had been altered to make it appear as if Levy Armstrong were openly weeping.

“I was there when they arrested her, and she definitely wasn’t crying — she was calm, rational, and dignified,” said Jordan Kushner, an attorney for Levy Armstrong. “This is part and parcel of a fascist regime where they literally invent reality to serve their fascist agenda.”

According to an initial Intercept analysis using Google SynthID — a program that identifies hidden markers used by Google AI tools on photos — the photo had been altered with the tech giant’s generative AI tools. Subsequent tests at Google’s request produced inconsistent results. Though Google’s tool initially stated that the crying image had been manipulated using its AI, a later test claimed that image was authentic. (A Google spokesperson said, “We’re trying to understand the discrepancy” — but did not answer repeated questions about it.)

Related

Trump’s War on America

In response to questions about the altered photo, a spokesperson for the White House referred The Intercept to a tweet from White House spokesperson Kaelan Dorr lashing out at “the people who feel the need to reflexively defend perpetrators of heinous crimes in our country.”

“Enforcement of the law will continue,” wrote Dorr. “The memes will continue.”

The original, unaltered image showing Levy Armstrong looking stalwart first appeared on the web in a pair of tweets by Homeland Security Secretary Kristi Noem, according to several image search engine tools.

About a half hour later, the White House posted its altered image showing Levy Armstrong in tears — including text labeling her as a “far-left agitator” and accusing her of “orchestrating church riots.”

The White House X account appears to have been the first place the altered image appeared on the web, according to the image search tools.

Related

Reproductive Rights Activists Charged Under Law Intended to Protect Abortion Clinics

Attorney General Pam Bondi announced Levy Armstrong’s arrest on Thursday. Along with Chauntyll Louisa Allen and William Kelly, Levy Armstrong faces charges under the Freedom of Access to Clinic Entrances Act, a 1994 law designed to limit anti-abortion protesters from impeding patients from seeking care. 

The arrests followed days of outrage online from the right over a protest on Sunday in which anti-ICE demonstrators entered the Cities Church, where a local ICE official serves as a pastor, according to The Associated Press.

“Religious freedom is the bedrock of this country,” Bondi wrote on X Thursday. “We will protect our pastors. We will protect our churches. We will protect Americans of faith.”

Jeffrey Lichtman, a defense attorney with numerous high-profile federal cases under his belt, told The Intercept that the post could conceivably have a prejudicial effect as the case against her proceeds.

“This altered photo makes her look weak and scared, and some people may interpret that as guilt,” Lichtman said. “I’d try to use it as evidence that this was a political prosecution. This isn’t, like, some aide that works in a congressional office somewhere, this is the White House, and it’s clear the White House controls Pam Bondi, and she’s the one responsible for this arrest.”

Ron Kuby, a veteran civil rights lawyer, told The Intercept that the problem lay less in the meme than in the prosecution itself. 

“As a defense lawyer, I’d work hard to make sure it wasn’t repeated, but it’s not going to result in dismissal of charges or any meaningful sanction from a judge,” Kuby said. “This is just Thursday in America. The outrage is not the graphic — the outrage is that they turned a simple disorderly conduct case into a federal prosecution for their propaganda efforts.”

Update: January 22, 2026, 5:27 p.m. ET
This story has been updated to reflect that Google declined to comment.

Update: January 24, 2026
This article and its headline were updated after subsequent tests of Google’s AI watermarking tool provided inconsistent responses about whether the image the White House posted on X had been manipulated by Google’s own AI tools.

The post White House Doctored Photo With AI to Make It Look Like an Activist Was Sobbing During Perp Walk appeared first on The Intercept.

This past July, YouTube deleted Inlakesh’s private backup account. And in August, Google, YouTube’s parent company, deleted his Google account, including his Gmail and his archive of documents and writings.

The tech giant initially claimed Inlakesh’s account violated YouTube’s community guidelines. Months later, the company justified his account termination by alleging his page contained spam or scam content.

However, when The Intercept inquired further about Inlakesh’s case, nearly two years after his account was deleted, YouTube provided a separate and wholly different explanation for the termination: a connection to an Iranian influence campaign.

YouTube declined to provide evidence to support this claim, stating that the company doesn’t discuss how it detects influence operations. Inlakesh remains unable to make new Google accounts, preventing him from sharing his video journalism on the largest English language video platform.

Inlakesh, now a freelance journalist, acknowledged that from 2019 to 2021 he worked from the London office of the Iranian state-owned media organization Press TV, which is under U.S. sanctions. Even so, Inlakesh said that should not have led to the erasure of his entire YouTube account, the vast majority of which was his own independent content that was posted before or after his time at Press TV.

A public Google document from the month Inlakesh’s account was deleted notes that the company had recently closed more than 30 accounts it alleged were linked to Iran that had posted content critical of Israel and its war on Gaza. The company did not respond when asked specifically if Inlakesh’s account was among those mentioned in the document.

Inlakesh said he felt like he was targeted not due to his former employer but because of his journalism about Palestine, especially amid the increasingly common trend of pro-Israeli censorship among Big Tech companies.

“What are the implications of this, not just for me, but for other journalists?” Inlakesh told The Intercept. “To do this and not to provide me with any information — you’re basically saying I’m a foreign agent of Iran for working with an outlet; that’s the implication. You have to provide some evidence for that. Where’s your documentation?”

Misdirection and Lack of Answers

Over the past couple years, YouTube and Google’s explanations given for the terminations of Inlakesh’s accounts have been inconsistent and vague.

YouTube first accused Inlakesh of “severe or repeated violations of our Community Guidelines.” When a Google employee, Marc Cohen, noticed Inlakesh’s public outcry about his account termination in February 2024, he decided to get involved. Cohen filed a support ticket on Google’s internal issue tracker system, “the Buganizer,” asking why a journalist’s account was deleted. Failing to get an answer internally, Cohen went public with his questions that March. After drawing the attention of the YouTube team on Twitter, he said he eventually received an internal response from Google which claimed that Inlakesh’s account had been terminated owing to “scam, deceptive or spam content.”

Cohen, who resigned from Google later that year over its support of the Israeli government’s genocide in Gaza, said had he not gotten involved, Inlakesh would have been left with even less information.

“They get away with that because they’re Google,” Cohen said. “What are you going to do? Go hire a lawyer and sue Google? You have no choice.”

When Inlakesh’s Gmail account was deleted this year, Google said his account had been “used to impersonate someone or misrepresent yourself,” which Google said is a violation of its policies. Inlakesh appealed three times but was given no response.

Only after The Intercept’s inquiry into Inlakesh’s case did Google shift its response to alleged Iranian influence.

“This creator’s channel was terminated in February 2024 as part of our ongoing investigations into coordinated influence operations backed by the Iranian state,” a YouTube spokesperson told The Intercept. The termination of his channel meant all other accounts associated with Inlakesh, including his backup account, were also deleted, YouTube said.

When The Intercept asked YouTube to elaborate on the reason behind the account deletions, such as which specific content may have flagged the account as being linked to an Iranian state influence operation, a YouTube spokesperson replied that YouTube doesn’t “disclose specifics of how we detect coordinated influence operations,” and instead referred The Intercept to Google’s Threat Analysis Group’s quarterly bulletins. TAG is a team within Google that describes itself as working “to counter government-backed hacking and attacks against Google and our users.”

Google’s Threat Analysis Group’s bulletin from when Inlakesh’s account was first terminated states that in February 2024, a total of 37 YouTube channels were deleted as a result of an “investigation into coordinated influence operations linked to Iran.” Four of these accounts, the document notes, were sharing content which “was critical of the Israeli government and its actions in the ongoing Israel-Gaza war” and had “shared content depicting alleged cyber attacks targeting Israeli organizations.” Google said in the document that the other 33 terminated YouTube channels had shown content “supportive of Iran, Yemen, and Palestine and critical of the US and Israel.”

A Pattern of Censorship

Google has a long-standing and well-documented practice of censoring Palestinian content or content critical of the Israeli government, in addition to evidence of human rights abuses in other conflicts. Such censorship has only exacerbated during Israel’s genocidal war on Gaza,

The company deploys various methods to censor content, such as teams of experts who manually review content, automated systems that flag content, reviews of U.S. sanction and foreign terror organization lists, as well as takedown requests from governments.

For the past decade, Israel’s Cyber Unit has openly run operations to convince companies to delete Palestine-related content from platforms such as YouTube.

Related

Israeli Group Claims It’s Working With Big Tech Insiders to Censor “Inflammatory” Wartime Content

Among U.S. allies, Israel had the highest percentage of requests resulting in takedowns on Google platforms, with a nearly 90 percent takedown rate, according to Google’s data since 2011. This rate outpaces countries like France, Germany, the United Kingdom, and Google’s home country, the United States. Absent from Google’s public reports, however, are takedown requests made by individual users, a route often weaponized by the Israeli cyber unit and internally by pro-Israel employees.

The scale of content deleted specifically due to U.S. sanctions is also difficult to quantify since such decisions happen without transparency. A recent investigation by The Intercept revealed that YouTube quietly deleted the accounts of three prominent Palestinian human rights organizations due to the Trump administration’s sanctions against the groups for assisting the International Criminal Court’s war crimes case against Israeli officials. The terminated pages accounted for at least 700 videos erased, many of which spotlighted alleged human rights abuses by the Israeli government.

Dia Kayyali, a technology and human rights consultant, said that in the past several years, as Big Tech platforms have relied more on automated systems that are fed U.S. sanction and terror lists, rights groups have seen an increase in the number of journalists within the Middle East and North Africa region who have had their content related to Palestine removed from YouTube, even when the content they post does not violate the company’s policies. The same could have happened with Inlakesh’s account, Kayyali said.

“And that’s part of the problem with automation — because it just does a really bad job of parsing content — content that could be graphic, anything that has any reference to Hamas,” Kayyali said. Hamas is included within the U.S. foreign terror organization list and Iran remains one of the most sanctioned countries by the U.S. government.

Google and other Big Tech platforms rely heavily on U.S. sanction lists in part to avoid potential liability from the State Department. But such caution is not always warranted, said Mohsen Farshneshani, principal attorney at the Washington, D.C.-based Sanctions Law Center.

Related

YouTube Quietly Erased More Than 700 Videos Documenting Israeli Human Rights Violations

Multinational corporations like Google tend to lean toward “overcompliance” with sanction regulations, often deleting content even when it legally is not required to do so, harming journalists and human rights groups, said Farshneshani.

Under U.S. law, in the Berman Amendment to the International Emergency Economic Powers Act, informational materials — in this case, reporting and journalism — are exempt from being subject to sanctions.

Such a carveout should have protected Inlakesh’s page from being deleted, Farshneshani said. Google likely could have taken down specific videos that raised concern, or demonetized specific videos or the entire account, he said. (Inlakesh said that years before terminating his videos and account, YouTube had demonetized some of his content depicting Israeli military violence.)

“Deleting an entire account is far from what the statutes or the regulations ask of U.S. entities,” Farshneshani said. “The exemption is meant for situations like this. And if these companies are to uphold their part of the bargain as brokers of information for the greater global community, they would do the extra leg work to make sure the stuff stays up.”

State-Sponsored Media

While YouTube and Google have not stated whether Inlakesh’s history with Press TV played a factor in the deletion, the Iranian state-funded outlet has long been under Google’s scrutiny. In 2013, Google temporarily deleted Press TV’s YouTube account before permanently deleting the channel in 2019 along with its Gmail account amid the first Trump administration’s sanctions campaign against Iran. The Biden administration in 2021 seized and censored dozens of websites tied to Iran, and in 2023 placed sanctions on Press TV due to Iran’s violent crackdown on anti-government protesters after the in-custody death of Mahsa Amini.

Press TV also has been accused by rights groups and journalists for filming and airing propaganda videos in which individuals detained by Iran are coerced to “confess” to alleged crimes in recorded interviews, as a part of the government’s attempts to justify their imprisonment or execution.

Press TV did not respond to The Intercept’s request for comment.

Out of the many videos on his YouTube account, Inlakesh recalled only two being associated with his work for Press TV: a documentary critical of the 2020 Trump deal on Israel–Palestine and a short clip about Republicans’ Islamophobic attacks on Rep. Ilhan Omar, D-Minn., in 2019. The rest either predate or postdate his stint at Press TV.

Press TV’s U.K. YouTube channel at times appears listed as an “associated channel” in archival versions of Inlakesh’s personal YouTube page. A YouTube spokesperson stated that YouTube uses “various signals to determine the relationship between channels linked by ownership for enforcement purposes,” but did not clarify what the specific signals were.

Inlakesh maintained that he had editorial independence while at Press TV and was never directed to post to his personal YouTube page.

Jillian York, the director for international freedom of expression at the Electronic Frontier Foundation, said she understood Google’s need to moderate content, but questioned why it deleted Inlakesh’s account rather than using its policy of labeling state-sponsored content, a system that itself has been plagued with problems. “More labels, more warnings, less censorship,” York said.

“The political climate around Palestine has made it such that a lot of the Silicon Valley-based social media platforms don’t seem particularly willing to ensure that Palestinian content can stay up,” she said.

Killing the Narrative

Inlakesh said he lost several documentaries about Israel and Palestine that were hosted exclusively on YouTube. However, what he lamented most was the loss of footage of his independent coverage from the West Bank, including livestreams that document alleged Israeli military abuses and were not backed up elsewhere.

One such video, he said, was a livestream from a protest at the major Israeli settlement of Beit El on February 11, 2020, against President Donald Trump’s lopsided annexation plan for Israel and Palestine.

Through the haze of tear gas, Inlakesh filmed Israeli soldiers camped out at a nearby hill, aiming their guns at the crowd of mostly children throwing rocks.

“And then you see the children drop,” Inlakesh recalled, followed by the bang of a gunshot. Paramedics rushed over to retrieve the children as Inlakesh followed behind. In all, Inlakesh said he filmed Israeli military gunfire hit three Palestinian children, a likely war crime violation, leaving them with wounds to the arms, legs and torso.

“You’re killing part of the narrative,” Inlakesh said. “You’re actively taking away the public’s ability to assess what happened at a critical moment during the history of the conflict.”

The post A Journalist Reported From Palestine. YouTube Deleted His Account Claiming He’s an Iranian Agent. appeared first on The Intercept.

Amandla Thomas-Johnson had attended a protest targeting companies that supplied weapons to Israel at a Cornell University job fair in 2024 for all of five minutes, but the action got him banned from campus. When President Donald Trump assumed office and issued a series of executive orders targeting students who protested in support of Palestinians, Thomas-Johnson and his friend Momodou Taal went into hiding.

Google informed Thomas-Johnson via a brief email in April that it had already shared his metadata with the Department of Homeland Security, as The Intercept previously reported. But the full extent of the information the agency sought — including usernames, addresses, itemized list of services, including any IP masking services, telephone or instrument numbers, subscriber numbers or identities, and credit card and bank account numbers — was not previously known.

“I’d already seen the subpoena request that Google and Meta had sent to Momodou [Taal], and I knew that he had gotten in touch with a lawyer and the lawyer successfully challenged that,” Thomas-Johnson said. “I was quite surprised to see that I didn’t have that opportunity.”

The subpoena provides no justification for why ICE is asking for this information, except that it’s required “in connection with an investigation or inquiry relating to the enforcement of U.S. immigration laws.” In the subpoena, ICE requests that Google not “disclose the existence of this summons for indefinite period of time.”

Thomas-Johnson, who is British, believes that ICE requested that information to track and eventually detain him — but he had already fled to Geneva, Switzerland, and is now in Dakar, Senegal. 

Related

Google Secretly Handed ICE Data About Pro-Palestine Student Activist

The Electronic Frontier Foundation, which is representing Thomas-Johnson, and the ACLU of Northern California sent a letter to Google, Amazon, Apple, Discord, Meta, Microsoft, and Reddit last week calling on tech companies to resist similar subpoenas in the future from DHS without court intervention. The letter asks the companies to provide users with as much notice as possible before complying with a subpoena to give them the opportunity to fight it, and to resist gag orders that would prevent the tech companies from informing targets that a subpoena was issued.

“Your promises to protect the privacy of users are being tested right now. As part of the federal government’s unprecedented campaign to target critics of its conduct and policies, agencies like DHS have repeatedly demanded access to the identities and information of people on your services,” the letter reads. “Based on our own contact with targeted users, we are deeply concerned your companies are failing to challenge unlawful surveillance and defend user privacy and speech.”

In addition to Thomas-Johnson’s case, the letter refers to other instances in which technology companies provided user data to DHS, including a subpoena sent to Meta to “unmask” the identities of users who documented immigration raids in California. Unlike Thomas-Johnson, users in that case were given the chance to fight the subpoena because they were made aware of it before Meta complied.

“Google has already fulfilled this subpoena,” an attorney for Google told Thomas-Johnson’s lawyer, as The Intercept previously reported. “Production consisted of basic subscriber information.”

The ICE subpoena requested the detailed information linked to Thomas-Johnson’s Gmail account. Thomas-Johnson confirmed to The Intercept that he had attached his bank and credit card numbers to his account to buy apps.

Google did not respond to a request for comment.

Lindsay Nash, a professor at Cardozo Law and a former staff attorney with ACLU Immigrants’ Rights Project, said that by not giving prior notice, Google deprived Thomas-Johnson of his ability to protect his information.

“The problem is that it doesn’t allow the person whose personal information is on the line and whose privacy may be being invaded to raise challenges to the disclosure of that potentially private information,” Nash said. “And I think that’s important to protect rights that they may have to their own information.”

Tech companies’ data sharing practices are primarily governed by two federal laws, the Stored Communications Act, which protects the privacy of digital communications, including emails, and Section 5 of the Federal Trade Commission Act, which prohibits unfair or deceptive trade practices.

“Under both federal law and the law of every state, you cannot deceive consumers,” said Neil Richards, a law professor at Washington University St. Louis who specializes in privacy, the internet, and civil liberties. “And if you make a material misrepresentation about your data practices, that’s a deceptive trade practice.”

Whether or not corporations are clear enough with consumers about how they collect and share their data has been litigated for decades, Richards said, referencing the infamous Cambridge Analytica lawsuit brought by the Federal Trade Commission, alleging that the company misled Facebook users about data collection and sharing.

Google’s public privacy policy acknowledges that it will share personal information in response to an “enforceable governmental request,” adding that its legal team will “frequently push back when a request appears to be overly broad or doesn’t follow the correct process.”

According to Google, the company overwhelmingly complied with the millions of requests made by the government for user information over the last decade. Its data also shows that those requests have spiked over the last five years. It’s unclear how many of those users were given notice of those requests ahead of time or after.

Richards said that cases like these emphasize the need for legal reforms around data privacy and urged Congress to amend the Stored Communications Act to require a higher standard before the government can access our digital data. He also said the federal government needs to regulate Big Tech and place “substantive restrictions on their ability to share information with the government.”

It’s hard to know exactly how tech companies are handling our personal data in relation to the government, but there seems to have been a shift in optics, Richards said. “What we have seen in the 12 months since the leaders of Big Tech were there on the podium at the inauguration,” Richards said, “is much more friendliness of Big Tech towards the government and towards state power.”

From Dakar, Thomas-Johnson said that understanding the extent of the subpoena was terrifying but had not changed his commitment to his work.

“As a journalist, what’s weird is that you’re so used to seeing things from the outside,” said Thomas-Johnson, whose work has appeared in outlets including Al Jazeera and The Guardian. “We need to think very hard about what resistance looks like under these conditions… where government and Big Tech know so much about us, can track us, can imprison, can destroy us in a variety of ways.”

Update: February 10, 5:54 p.m. ET

This story has been updated to reflect that Thomas-Johnson’s legal team still does not know the full extent of the information that Google provided to ICE, but that Thomas-Johnson said his bank and credit card numbers were attached to his account.

The post Google Fulfilled ICE Subpoena Demanding Student Journalist’s Bank and Credit Card Numbers appeared first on The Intercept.

The documents, many heavily redacted, consist of contract paperwork from deals originally signed in July 2025 between the Department of Defense and Anthropic, Google, OpenAI, and xAI, as well as later amendments to those agreements. Through the contracts, each worth up to $200 million, the companies agreed to build prototypes of artificial intelligence tools to “improve military advantage, military utility, or enhance military decision making” across the whole of the armed forces. These prototypes, aimed at applications from “warfighting” and automated decision-making to logistics and military intelligence, would go on to form the basis of the artificial intelligence tools now deployed across the Department of Defense’s classified networks. The document set includes one contract reflecting OpenAI’s classified network deployment; The Intercept is awaiting the government’s production of equivalent documents from the other contractors.

Related

The Pentagon Asked OpenAI for Artificial Intelligence Designed to Rarely Say No

The contract materials show the close and complex partnership between the Pentagon and the AI industry, and detail the extent to which AI firms themselves are expected to influence how their technology is used to augment the U.S. capacity to wage war.

“This appears to be the first time the extent of collaboration between frontier AI labs and the Pentagon is spelled out in explicit terms,” Sophia Goodfriend, a University of Cambridge research fellow and non-resident fellow at the Harvard Kennedy School’s Middle East Initiative studying the impact of big data and machine learning on military conflict, told The Intercept. “What’s particularly notable is the terms by which engineers are working in lockstep with the department of war to engineer AI systems for surveillance, targeting, and killing.”

Official announcements following the 2025 deals were light on details and heavy on platitudes. “We must equip our warfighters with 21st-century technology in order to defeat 21st-century threats,” explained Pentagon spokesperson Sean Parnell at the time.

Google was similarly cagey: “These advanced AI solutions will enable the DoD to effectively address defense challenges and scale the adoption of agentic AI across enterprise systems to drive innovation and efficiency with agile, proven technology,” the company said.

The contract documents for the prototyping project paint a clearer picture, revealing deals that involve far more than a simple sale of software.

An “information sharing” provision in all four agreements, for instance, creates a two-way exchange of data and expertise. The companies agreed not only to provide access to their best large language models, but also to “inform DOD and CDAO [Chief Digital and Artificial Intelligence Office] strategy on frontier AI and AI,” as phrased in the agreement with Google.

Under the agreements, the AI labs were approved to receive an array of sensitive information, which could include “benchmarks datasets for DoD use cases, appropriate briefings on DoD operational missions and threats to inform development / release of technologies, their application to DoD operational problems, or DoD plans and strategies for future AI adoption.”

All four companies would in turn provide the Pentagon “feedback on DoD strategies for frontier AI adoption to ensure that AI capabilities can be delivered and scaled to meet the warfighter’s needs.” The companies also pledged to offer briefings on corporate inner workings, including “frontier AI model performance, case studies on current or likely future frontier AI applications, or projections of future trends in frontier AI maturation.”

Related

Anthropic Says We Must Stop Authoritarian AI. But What About Its Authoritarian Investors?

The labs would sometimes be the ones offering intelligence briefings on foreign actors, with Google tapped to inform the Pentagon on “AI tactics and techniques of adversaries” of the United States. These briefings would be conducted in both unclassified and classified settings.

The collaboration also entails Big Tech providing training and educational seminars directly to the military, including “Presentations and briefings at technical exchange meetings on relevant topics, including frontier AI model performance and case studies of frontier AI applications.” One document notes that Google, for example, would lead multiple seminars on “responsible AI” deployment.” The documents state engineers and policy experts from all four contractors will conduct “tabletop exercises,” gamified simulations of real-world scenarios, with the Pentagon.

This back and forth between the military and private sector would, the Pentagon hoped, allow for “iterative refinement of frontier AI models to address real-world challenges in areas such as intelligence analysis, cybersecurity, and autonomous systems,” according to contract language with all four companies.

The AI companies were contractually scheduled to provide “risk forecasting, and threat ideation exercises” — predictions of pitfalls and dangers their own products might pose in the future. “Frontier AI labs have advanced risk forecasting tradecraft and a clear vision of the next wave of frontier AI technological developments,” the contracts read. “Frontier AI companies can help DoD gain a better understanding of the risks that might be associated with subsequent near-term frontier AI models or features, to avoid strategic surprise and to mitigate the associated risks.”

Heidy Khlaaf, chief scientist at the AI Now Institute and former safety engineer at OpenAI, told The Intercept this is a dubious claim with dangerous implications. Pointing to recent disclosures by OpenAI and Anthropic that semi-autonomous large language models broke into computer networks owned by other companies during testing, Khlaaf questioned their fitness to help build guardrails in the first place. “This is a very concerning development,” she said.

According to Khlaaf, the “risk forecasts” so far offered by frontier AI labs have been light on evidence and tend to “emphasize self-beneficial skewed risks such as a purported AI arms race and speculative ‘existential’ risks,” like self-aggrandizing science fiction-style “Terminator” scenarios.

The public would be better served, she said, with a reliance on independent assessment of the risks these companies present, not the word of the companies themselves. Trusting companies to self-report the risks of their own products constitutes both a conflict of interest, Khlaaf added, and a “subversion of democratic processes when AI labs are allowed to take over the arbitration of risk determinations with life-or-death consequences.”

As the prototyping effort progressed, the contract documents show the scope of the project expanded. New versions of the contract show additional deliverables included in the list of what the Pentagon was purchasing. The government redacted descriptions of these new deliverables on the basis of military secrecy and corporate confidentiality.

It was during the negotiation of these subsequent amendments when Anthropic and the Pentagon had their falling out. Anthropic refused to sign a follow-up deal to allow deployment of its technology on classified military networks without contractual prohibitions against use for domestic spying and autonomous weapons, leading Secretary of Defense Pete Hegseth to designate the firm a “supply chain risk” in March and ban its services from government use, a decision overturned by a federal judge last month.

Related

OpenAI on Surveillance and Autonomous Killings: You’re Going to Have to Trust Us

The Department of Defense quickly reached follow-up agreements with Anthropic’s main competitors — Google, OpenAI, and xAI — to apply the tools they had been developing since the summer of 2025 to classified military networks. OpenAI’s deal, described by CEO Sam Altman as “definitely rushed,” was reportedly finalized on February 27. The following day, the company published a blog post defending “Our agreement with the Department of War” and insisting there were clauses prohibiting its technology’s use for domestic surveillance.

An amended contract document obtained by The Intercept through its lawsuit dated February 27 contains new language, including a clause noting that OpenAI will embed company engineers within the military to aid U.S. Government employees in implanting and utilizing its AI system. These workers, the document notes, “can be deployed to warfighting support settings including, but not limited to, combatant commands, service components, and theatre components.” The February 27 amendments also include a section on “System Oversight” is redacted in its entirety.

None of the documents released to The Intercept pertaining to xAI or Google’s equivalent agreements contain any such oversight section.

Two documents indicate the Department of Defense asked OpenAI to minimize the extent to which its large language model would refuse to do what was requested by the military. The Department of Defense and OpenAI said those documents were draft materials incorrectly provided to The Intercept, and that the final contract included no such provision.

(In 2024, The Intercept sued OpenAI in federal court over the company’s use of copyrighted articles to train its chatbot ChatGPT. The case is ongoing.)

Related

OpenAI Quietly Deletes Ban on Using ChatGPT for “Military and Warfare”

The contractual coziness with the world’s most powerful military indicates just how far the artificial intelligence industry has strayed from its earlier public commitments to safety and global prosperity. OpenAI, originally founded as a nonprofit “to ensure that artificial general intelligence benefits all of humanity,” up until two years ago explicitly banned “military and warfare” use in its terms of service. The following year Google reversed a similar self-imposed ban on using its state-of-the-art AI services for surveillance or “technologies whose principal purpose or implementation is to cause or directly facilitate injury to people.”

Both companies are now eagerly landing slices of the nearly trillion-dollar Pentagon budget, and for opportunities to augment the killing power of a military that has recently bombed schoolgirls in Iran and civilian vessels in the Caribbean.

Goodfriend, the researcher, said the contracts illustrate the illusory nature of values statements published by these labs. “In recent months companies like Anthropic and Open AI have maintained ‘firm red lines’ when it comes to developing autonomous weapons systems and mass surveillance tools” long after they signed contracts to build “autonomous systems” for the military, she said.

“All this indicates these labs and their personnel have been far more closely integrated into DOD AI development than previously understood and casts some doubt on the supposed contractual limits they have intended to place on the DoD’s use of these systems,” Goodfriend said.

Anthropic, Google, and xAI did not provide comment or answer questions about their respective contracts. In a statement to The Intercept, OpenAI spokesperson Nate Evans said “Our tools are built to support legitimate national security work — including vital cyber defense efforts — while refusing uses that cross the red lines we publicly announced. These restrictions are also reflected in our signed agreement: no use of OpenAI technology for mass domestic surveillance, to direct autonomous weapons systems, or for high-stakes automated decisions. We will continue to work closely with the Department to make sure our AI tools are deployed responsibly and safely.”

Related

The Dawn of a New Cold War

On the long timeline of American warfare, these artificial intelligence tools are in their infancy. The Pentagon’s use of artificial intelligence in the process of killing, however, is not a hypothetical. In March, the Wall Street Journal reported that despite Hegseth’s ban, U.S. Central Command used Anthropic’s large language model services to execute its bombardment of Iran, including for “target identification.”

Despite the intimacy of technology firms and military using it to pinpoint airstrikes, it’s unclear what oversight, if any, these companies have over their own weaponized software once it is in the Pentagon’s hands. When asked if his company’s tools were used in an American airstrike on the first day of the Iran war that killed 120 Iranian schoolchildren, Anthropic CEO Dario Amodei told Bloomberg News he simply did not know. An August 24 report from the New York Times detailed the use of fully autonomous drones guided by machine learning software running on chips made by Nvidia, whose processors have helped make the explosive growth of Anthropic and OpenAI possible. The full extent of the U.S. tech sector’s involvement in killing abroad remains obscured behind trade secrecy and national security considerations.

“The same technology we saw autonomously hacking websites on the open internet weeks ago is being used to inform DoD decision-making,” said Vivian Dong of Legal Advocates for Safe Science and Technology. “The public deserves insight into how and under what constraints the Department is using this technology.”

The post AI Giants Work Hand-in-Hand With the Pentagon, Contracts Reveal appeared first on The Intercept.

As they feared, Homeland Security Investigations, the intelligence division of U.S. Immigration and Customs Enforcement, was intent to track them both down. As agents scrambled to find Taal and Thomas-Johnson, HSI sent subpoenas to Google and Meta for sensitive data information about their Gmail, Facebook, and Instagram accounts.

In Thomas-Johnson’s case, The Intercept found, Google handed over data to ICE before notifying him or giving him an opportunity to challenge the subpoena. By the time he found out about the data demand, Thomas-Johnson had already left the U.S.

During the first Trump administration, tech companies publicly fought federal subpoenas on behalf of their users who were targeted for protected speech — sometimes with great fanfare. With ICE ramping up its use of dragnet tools to meet its deportation quotas and smoke out noncitizens who protest Israel’s war on Gaza, Silicon Valley’s willingness to accommodate these kinds of subpoenas puts those who speak out at greater risk.

Lindsay Nash, a professor at Cardozo School of Law in New York who has studied ICE’s use of administrative subpoenas, said she was concerned but not surprised that Google complied with the subpoena about Thomas-Johnson’s account without notifying him.

“Subpoenas can easily be used and the person never knows,” Nash told The Intercept. “It’s problematic to have a situation in which people who are targeted by these subpoenas don’t have an opportunity to vindicate their rights.”

Google declined to discuss the specifics of the subpoenas, but the company said administrative subpoenas like these do not include facts about the underlying investigation.

“Our processes for handling law enforcement subpoenas are designed to protect users’ privacy while meeting our legal obligations,” said a Google spokesperson in an emailed statement. “We review every subpoena and similar order for legal validity, and we push back against those that are overbroad or improper, including objecting to some entirely.”

ICE agents sent the administrative subpoenas to Google and Meta by invoking a broad legal provision that gives immigration officers authority to demand documents “relating to the privilege of any person to enter, reenter, reside in, or pass through the United States.”

One recent study based on ICE records found agents invoke this same provision hundreds of times each year in administrative subpoenas to tech companies. Another study found ICE’s subpoenas to tech companies and other private entities “overwhelmingly sought information that could be used to locate ICE’s targets.”

Unlike search warrants, administrative subpoenas like these do not require a judge’s signature or probable cause of a crime, which means they are ripe for abuse.

HSI had flagged Taal to the State Department following “targeted analysis to substantiate aliens’ alleged engagement of antisemitic activities,” according to an affidavit later filed in court by a high-ranking official. This analysis amounted to a trawl of online articles about Taal’s participation in Gaza protests and run-ins with the Cornell administration. The State Department revoked Taal’s visa, and ICE agents in upstate New York began searching for him.

In mid-March, the week after Mahmoud Khalil was arrested in New York City, Taal sued the Trump administration, seeking an injunction that would have blocked ICE from detaining him too. By this point, he and Thomas-Johnson had both left their campus housing at Cornell and were hiding from ICE in a house 10 miles outside Ithaca.

Two days after Taal filed his suit, still unable to track him down, ICE sent an administrative subpoena to Meta. According to notices Meta emailed to Taal, the subpoena sought information about his Instagram and Facebook accounts. Meta gave Taal 10 days to challenge the subpoena in court before the company would comply and hand over data about his accounts to ICE.

Like Google, Meta declined to discuss the subpoena it received about Taal’s account, referring The Intercept to a webpage about the company’s compliance with data demands.

A week later, HSI sent another administrative subpoena to Google regarding Taal’s Gmail account, according to a notice Google sent him the next day.

“It was a phishing expedition,” Taal said in a text message to The Intercept.

After Taal decided to leave the country and dismissed his lawsuit in April, ICE withdrew its subpoenas for his records.

Do you have information about DHS or ICE targeting activists online? Use a personal device to contact Shawn Musgrave on Signal at shawnmusgrave.82

But on the last day of March, HSI sent yet another subpoena, this one to Google for information about Thomas-Johnson’s Gmail account. Without giving Thomas-Johnson any advance warning or the opportunity to challenge it, Google complied with the subpoena, and it only notified him weeks later. 

“Google has received and responded to legal process from a Law Enforcement authority compelling the release of information related to your Google Account,” read an email Google sent him in early May.

By this point, Thomas-Johnson had already left the country too. He fled after a friend was detained at the Tampa airport, handed a note with Thomas-Johnson’s name on it, and asked repeatedly about his whereabouts, he told The Intercept.

Thomas-Johnson’s lawyer, who also represented Taal, reached out to an attorney for Google about the demand for his client’s account information.

“Google has already fulfilled this subpoena,” Google’s attorney replied by email, further explaining that Google’s “production consisted of basic subscriber information,” such as the name, address, and phone number associated with the account. Google did not produce “the contents of communications, metadata regarding those communications, or location information,” the company’s attorney wrote.

“This is the extent that they will go to be in support of genocide,” Taal said of the government’s attempts to locate him using subpoenas.

Correction: September 16, 2025, 12:40 p.m. ET
The story has been updated to correct the spelling of Amandla Thomas-Johnson’s last name.

The post Google Secretly Handed ICE Data About Pro-Palestine Student Activist appeared first on The Intercept.

In an era when companies known for e-commerce, social networks, and search engines have also become close collaborators with militaries, is bombing their servers fair game?

Three days after the U.S. and Israel began their joint bombardment, the Islamic Revolutionary Guard Corps launched kamikaze drone strikes against Amazon-owned data centers in the United Arab Emirates and Bahrain that provide an array of cloud computing services to customers throughout the Middle East. The impacts and subsequent fires “caused structural damage, disrupted power delivery to our infrastructure, and in some cases required fire suppression activities that resulted in additional water damage,” according to Amazon, resulting in service outages across the region.

The motive behind the attack, according to Iranian state television, was not to block people from ordering groceries or posting to social media, but rather to highlight “the role of these centers in supporting the enemy’s military and intelligence activities.” Though only Amazon’s centers are known to have come under fire, a March 11 tweet from the quasi-official Tasnim News Agency listed dozens of regional facilities, including data centers owned by Microsoft, Google and others, deemed “Enemy Technology Infrastructure” suitable for targeting.

It’s unclear if the Amazon data centers struck by Iranian drone strikes are used for military purposes or civilian purposes, or both. And it’s unknown if the attacks in any way hindered the militaries of the U.S., Israel, or their allies in the Gulf from using AI or other cloud-based services in their war efforts. But with Amazon, Google, and even Facebook parent company Meta are all eager partners of the Pentagon that augment the destructive power of the United States in Iran and elsewhere, server farms may now have the same status as factories building bombs and warplanes.

Scholars of international law and the laws of armed conflict say that when a military runs on the cloud, the cloud becomes a legal military target. But the cloud is an abstraction, not a physical site — a global network of millions of chips in servers spread across hundreds of massive buildings across the planet, servicing both civilian apps and state tools used to surveil and kill. Separating the former from the latter is an extremely difficult task.

“The legality turns on whether the specific facility, at the specific moment, is genuinely serving the military operations of a party to the conflict in a way that offers a concrete and definite advantage to the attacker,” explained León Castellanos-Jankiewicz, a lawyer with the Asser Institute for International and European Law in The Hague.

Sometimes the split between military and civilian use is straightforward. Microsoft, for example, helps run the Joint Warfighter Cloud Capability, which the Pentagon says provides it with “greater lethality.” This work involves the processing of classified data, which the government does not want commingling with civilian tech. Cloud computing services are generally offered via geographically distinct “regions,” each made up of many physical data centers. Customers typically select the region that is closest to them to minimize lag time. Microsoft’s US DoD Central and US DoD East regions are “reserved for exclusive [Department of Defense] use,” according to the company, and are serviced by data centers in Des Moines, Iowa, and Northern Virginia, respectively.

Amazon offers similar cloud regions exclusive for Pentagon use, though the location of these data centers is not public. Oracle, another JWCC provider, operates Pentagon-specific facilities in Chicago, Phoenix, and Virginia. Companies are understandably tight-lipped about where exactly on the map these facilities stand, in no small part because Iran, or any country at war with the U.S., would have reason to target them.

“A data center that is used solely or primarily for military applications is targetable,” said Ioannis Kalpouzos, an international law scholar and visiting professor at Harvard Law, “and a center that supports the Pentagon’s JWCC falls in that category.”

Related

AI’s Imperial Agenda

The march of data center construction has become a point of contention across the United States and around the world, with communities frequently — and sometimes successfully — rallying to block what they view as enormous resource-draining eyesores. But for those living in the widening shadow of data centers, planned or built, their status as military targets may be unsettling beyond concerns over water and energy consumption.

And as Defense Secretary Pete Hegseth aggressively shoehorns AI tools into the military wherever possible, the rapid expansion of data centers means the potential proliferation of legitimate military targets across the United States.

With comparisons between the destructive power of AI-augmented warfare and nuclear weaponry becoming more common, the ever-expanding network of American data centers may recreate Cold War anxieties around intercontinental ballistic missile, or ICBM, silo placement. The country’s nuclear launch capabilities were famously clustered in the relatively sparsely populated Upper Midwest, forming a so-called “nuclear sponge” that would draw Soviet nukes away from population centers and toward rural areas and farmland.

But the legal calculus around most data centers will be less clear. Google, for example, says the Pentagon uses both its general purpose public cloud and smaller specialized air-gapped networks that don’t touch the public internet, depending on the sensitivity of the data involved. Even cloud work involving Top Secret military data “can operate within Google’s trusted, secure, and managed data centers.” The company also sells modular mini-data centers for use closer to battlefields or bases.

These arrangements, shrouded in both military and trade secrecy, make it hard to assess whether a server is hosting a student’s homework or Air Force R&D, blurring the legality of attacking data centers that may host both. Google may have little control over how governments use its cloud tools; The Intercept has previously reported that Google executives worried internally they wouldn’t be able to tell how the Israeli military was deploying its cloud services.

“The practical challenge is that cloud infrastructure is often technically opaque, even to providers themselves,” Castellanos-Jankiewicz said. “The services a given data center supports may not be readily ascertainable from the outside or even inside, which complicates the attacker’s legal obligations considerably.”

Amazon and Google’s Project Nimbus similarly provides cloud computing services across the Israeli government, including both civilian agencies and the Ministry of Defense, along with state-owned weapons companies.

“The picture becomes more legally complex when a data center functions as a so-called ‘dual-use’ object,” simultaneously hosting military data or capabilities alongside civilian services,” Castellanos-Jankiewicz told The Intercept. “Once a facility is found to make an effective contribution to military action, the entire physical object can, under the dominant legal view, qualify as a military objective.”

The embrace of commercial cloud computing by the U.S. and others has muddled an already murky legal picture, Castellanos-Jankiewicz explained. “A military’s decision to store classified data or run AI-enabled military systems on commercial cloud infrastructure shared with civilian services could itself raise legal concerns — particularly if the commingling of military and civilian uses makes a strike more likely or increases the foreseeable harm to civilians when one occurs.”

Related

OpenAI on Surveillance and Autonomous Killings: You’re Going to Have to Trust Us

Determining whether a given data center can be legally attacked under international humanitarian law — itself comprised of various treaties that not every country adheres to — relies on a complex series of balancing tests that rarely produce concrete answers. To begin with, every object and person is generally presumed civilian and exempt from attack under this framework. Before launching a strike, a country is supposed to have a verifiable reason to believe a data center contributes to the enemy war effort, and reason to believe an attack will appreciably harm that effort. What “effectively contributes to military action” will, of course, be a source of disagreement.

Anthropic’s Claude large language model was reportedly used to accelerate American airstrikes against Iran; Claude, in turn, was built in part using 500,000 chips housed in an $11 billion Amazon data center in Indiana. If Claude is now arguably a weapon, is this Indiana site the data equivalent of a bomb factory? Kalpouzos, the Harvard Law visiting professor, told The Intercept it depends on the facts at the moment the bomb hits, not past usage. “If the facility is currently used in the training of the LLM that is used in the conduct of military operations — for example, by fine-tuning object classification or user-interaction features — then this could render it targetable,” he said.

In a recent article for Just Security, Klaudia Klonowska and Michael Schmitt said that the law calls for proportionality and restraint even against military targets. An attack against a data center that provided both military and civilian computing would need to be precise enough to destroy the former while minimizing harm to the latter, they argued. But international law may call for a degree of carefulness that militaries have little interest in. “If it were possible to attack only the area of the data center where servers hosting military data are located without destroying the entire center, the attacker would need to do so,” they wrote.

These requirements can be hard to observe in reality. The U.S. and Israel both tout the extreme precision of airstrikes that regularly slaughter civilians. And neither country, nor Iran, is a signatory to some of the relevant legal frameworks that make up the so-called “laws of armed conflict” in the first place.

Indiscriminate warfare practice by U.S. and Israel has also, ironically, been instrumental in reshaping how these laws are interpreted and effectively loosened. Throughout the Israeli genocide in Gaza, Israel’s military and the Pentagon both made clear it’s acceptable to destroy an apartment block or hospital if one first claims there is a genuine military target inside.

The second Trump administration in particular has been keen to more tightly integrate Silicon Valley into the global American killing apparatus, a plan to which the industry has shown itself to be largely amenable. Even after being thoroughly maligned by the administration following the collapse of its Pentagon deal over purported disagreements around safety guardrails, Anthropic CEO Dario Amodei issued a public statement making clear he still wanted in on military spending: “Anthropic has much more in common with the Department of War than we have differences. We both are committed to advancing US national security and defending the American people, and agree on the urgency of applying AI across the government.” That attitude, now commonplace across the tech sector, will see the further commingling of consumer tech and warfare both in the abstract and under sprawling data center rooftops across the country.

“These [data centers] are further melding military and civilian infrastructure,” said Kalpouzos, “and together with the increasingly permissive rules of engagement adopted by the U.S. and Israel, are potentially drawing in larger sectors of the economy and society in what is targeted and destroyed.”

The post Data Centers Are Military Targets Now appeared first on The Intercept.

The report makes explicit the extent to which the tech giant understood the risk of providing state-of-the-art cloud and machine learning tools to a nation long accused of systemic human rights violations and wartime atrocities. Not only would Google be unable to fully monitor or prevent Israel from using its software to harm Palestinians, but the report also notes that the contract could obligate Google to stonewall criminal investigations by other nations into Israel’s use of its technology. And it would require close collaboration with the Israeli security establishment — including joint drills and intelligence sharing — that was unprecedented in Google’s deals with other nations.

A third-party consultant Google hired to vet the deal recommended that the company withhold machine learning and artificial intelligence tools from Israel because of these risk factors.

Three international law experts who spoke with The Intercept said that Google’s awareness of the risks and foreknowledge that it could not conduct standard due diligence may pose legal liability for the company. The rarely discussed question of legal culpability has grown in significance as Israel enters the third year of what has widely been acknowledged as a genocide in Gaza — with shareholders pressing the company to conduct due diligence on whether its technology contributes to human rights abuses.

“They’re aware of the risk that their products might be used for rights violations,” said León Castellanos-Jankiewicz, a lawyer with the Asser Institute for International and European Law in The Hague, who reviewed portions of the report. “At the same time, they will have limited ability to identify and ultimately mitigate these risks.”

Google declined to answer any of a list of detailed questions sent by The Intercept about the company’s visibility into Israel’s use of its services or what control it has over Project Nimbus.

Company spokesperson Denise Duffy-Parkes instead responded with a verbatim copy of a statement that Google provided for a different article last year. “We’ve been very clear about the Nimbus contract, what it’s directed to, and the Terms of Service and Acceptable Use Policy that govern it. Nothing has changed.”

Portions of the internal document were first reported by the New York Times, but Google’s acknowledged inability to oversee Israel’s usage of its tools has not previously been disclosed.

In January 2021, just three months before Google won the Nimbus contract alongside Amazon, the company’s cloud computing executives faced a dilemma.

The Project Nimbus contract — then code-named “Selenite” at Google — was a clear moneymaker. According to the report, which provides an assessment of the risks and rewards of this venture, Google estimated a bespoke cloud data center for Israel, subject to Israeli sovereignty and law, could reap $3.3 billion between 2023 and 2027, not only by selling to Israel’s military but also its financial sector and corporations like pharmaceutical giant Teva.

But given decades of transgressions against international law by Israeli military and intelligence forces it was now supplying, the company acknowledged that the deal was not without peril. “Google Cloud Services could be used for, or linked to, the facilitation of human rights violations, including Israeli activity in the West Bank,” resulting in “reputation harm,” the company warned.

In the report, Google acknowledged the urgency of mitigating these risks, both to the human rights of Palestinians and Google’s public image, through due diligence and enforcement of the company’s terms of service, which forbid certain acts of destruction and criminality.

But the report makes clear a profound obstacle to any attempt at oversight: The Project Nimbus contract is written in such a way that Google would be largely kept in the dark about what exactly its customer was up to, and should any abuses ever come to light, obstructed from doing anything about them.

The document lays out the limitations in stark terms.

Google would only be given “very limited visibility” into how its software would be used. The company was “not permitted to restrict the types of services and information that the Government (including the Ministry of Defense and Israeli Security Agency) chooses to migrate” to the cloud.

Attempts to prevent Israeli military or spy agencies from using Google Cloud in ways damaging to Google “may be constrained by the terms of the tender, as Customers are entitled to use services for any reason except violation of applicable law to the Customer,” the document says. A later section of the report notes Project Nimbus would be under the exclusive legal jurisdiction of Israel, which, like the United States, is not a party to the Rome Statute and does not recognize the International Criminal Court.

Should Project Nimbus fall under legal scrutiny outside of Israel, Google is required to notify the Israeli government as early as possible, and must “Reject, Appeal, and Resist Foreign Government Access Requests.”

Google noted this could put the company at odds with foreign governments should they attempt to investigate Project Nimbus. The contract requires Google to “implement bespoke and strict processes to protect sensitive Government data,” according to a subsequent internal report, also viewed by The Intercept that was drafted after the company won its bid. This obligation would stand even if it means violating the law: “Google must not respond to law enforcement disclosure requests without consultation and in some cases approval from the Israeli authorities, which could cause us to breach international legal orders / law.”

The second report notes another onerous condition of the Nimbus deal: Israel “can extend the contract up to 23 years, with limited ability for Google to walk away.”

The initial report notes that Google Cloud chief Thomas Kurian would personally approve the contract with full understanding and acceptance of these risks before the company submitted its contract proposal. Google did not make Kurian available for comment.

Business for Social Responsibility, a human rights consultancy tapped by Google to vet the deal, recommended the company withhold machine learning and AI technologies specifically from the Israeli military in order to reduce potential harms, the document notes. It’s unclear how the company could have heeded this advice considering the limitations in the contract. The Intercept in 2022 reported that Google Cloud’s full suite of AI tools was made available to Israeli state customers, including the Ministry of Defense.

BSR did not respond to a request for comment.

The first internal Google report makes clear that the company worried how Israel might use its technology. “If Google Cloud moves forward with the tender, we recommend the business secure additional assurances to avoid Google Cloud services being used for, or linked to, the facilitation of human rights violations.”

It’s unclear if such assurances were ever offered.

Related

U.S. Companies Honed Their Surveillance Tech in Israel. Now It’s Coming Home.

Google has long defended Project Nimbus by stating that the contract “is not directed at highly sensitive, classified or military workloads relevant to weapons or intelligence services.” The internal materials note that Project Nimbus will entail nonclassified workloads from both the Ministry of Defense and Shin Bet, the country’s rough equivalent of the FBI. Classified workloads, one report states, will be handled by a second, separate contract code-named “Natrolite.” Google did not respond when asked about its involvement in the classified Natrolite project.

Both documents spell out that Project Nimbus entails a deep collaboration between Google and the Israeli security state through the creation of a Classified Team within Google. This team is made up of Israeli nationals within the company with security clearances, designed to “receive information by [Israel] that cannot be shared with [Google].” Google’s Classified Team “will participate in specialized training with government security agencies,” the first report states, as well as “joint drills and scenarios tailored to specific threats.”

The level of cooperation between Google and the Israeli security state appears to have been unprecedented at the time of the report. “The sensitivity of the information shared, and general working model for providing it to a government agency, is not currently provided to any country by GCP,” the first document says.

Whether Google could ever pull the plug on Nimbus for violating the company rules or the law is unclear. The company has claimed to The Intercept and other outlets that Project Nimbus is subject to its standard terms of use, like any other Google Cloud customer. But Israeli government documents contradict this, showing the use of Project Nimbus services is constrained not by Google’s normal terms, but a secret amended policy.

A spokesperson for the Israeli Ministry of Finance confirmed to The Intercept that the amended Project Nimbus terms of use are confidential. Shortly after Google won the Nimbus contract, an attorney from the Israeli Ministry of Finance, which oversaw the deal, was asked by reporters if the company could ever terminate service to the government. “According to the tender requirements, the answer is no,” he replied.

In its statement, Google points to a separate set of rules, its Acceptable Use Policy, that it says Israel must abide by. These rules prohibit actions that “violate or encourage the violation of the legal rights of others.” But the follow-up internal report suggests this Acceptable Use Policy is geared toward blocking illegal content like sexual imagery or computer viruses, not thwarting human rights abuses. Before the government agreed to abide by the AUP, Google wrote there was a “relatively low risk” of Israel violating the policy “as the Israel government should not be posting harmful content itself.” The second internal report also says that “if there is a conflict between Google’s terms” and the government’s requirements, “which are extensive and often ambiguous,” then “they will be interpreted in the way which is the most advantageous to the customer.”

International law is murky when it comes to the liability Google could face for supplying software to a government widely accused of committing a genocide and responsible for the occupation of the West Bank that is near-universally considered illegal.

Related

Google Won’t Say Anything About Israel Using Its Photo Software to Create Gaza “Hit List”

Legal culpability grows more ambiguous the farther you get from the actual act of killing. Google doesn’t furnish weapons to the military, but it provides computing services that allow the military to function — its ultimate function being, of course, the lethal use of those weapons. Under international law, only countries, not corporations, have binding human rights obligations. But if Project Nimbus were to be tied directly to the facilitation of a war crime or other crime against humanity, Google executives could hypothetically face criminal liability under customary international law or through a body like the ICC, which has jurisdiction in both the West Bank and Gaza.

Civil lawsuits are another option: Castellanos-Jankiewicz imagined a scenario in which a hypothetical plaintiff with access to the U.S. court system could sue Google over Project Nimbus for monetary damages, for example.

Along with its work for the Israeli military, Google through Project Nimbus sells cloud services to Israel Aerospace Industries, the state-owned weapons maker whose munitions have helped devastate Gaza. Another confirmed Project Nimbus customer is the Israel Land Authority, a state agency that among other responsibilities distributes parcels of land in the illegally annexed and occupied West Bank.

An October 2024 judicial opinion issued by the International Court of Justice, which arbitrates disputes between United Nations member states, urged countries to “take all reasonable measures” to prevent corporations from doing anything that might aid the illegal occupation of the West Bank. While nonbinding, “The advisory opinions of the International Court of Justice are generally perceived to be quite authoritative,” Ioannis Kalpouzos, a visiting professor at Harvard Law and expert on human rights law and laws of war, told The Intercept.

Establishing Google’s legal culpability in connection with the occupation of the West Bank or ongoing killing in Gaza entails a complex legal calculus, experts explained, hinging on the extent of its knowledge about how its products would be used (or abused), the foreseeability of crimes facilitated by those products, and how directly they contributed to the perpetration of the crimes. “Both the very existence of the document and the language used suggest at least the awareness of the likelihood of violations,” Kalpouzos said.

While there have been a few instances of corporate executives facing local criminal charges in connections with human rights atrocities, liability stemming from a civil lawsuit is more likely, said Castellanos-Jankiewicz. A hypothetical plaintiff might have a case if they could demonstrate that “Google knew or should have known that there was a risk that this software was going to be used or is being used,” he explained, “in the commission of serious human rights violations, war crimes, crimes against humanity, or genocide.”

Getting their date in court before an American judge, however, would be another matter. The 1789 Alien Tort Statute allows federal courts in the United States to take on lawsuits by foreign nationals regarding alleged violations of international law but has been narrowed considerably over the years, and whether U.S. corporations could even be sued under the statute in the first place remains undecided.

History has seen scant few examples of corporate accountability in connection with crimes against humanity. In 2004, IBM Germany donated $4 million to a Holocaust reparations fund in connection with its wartime role supplying computing services to the Third Reich. In the early 2000s, plaintiffs in the U.S. sued dozens of multinational corporations for their work with apartheid South Africa, including the sale of ”essential tools and services,” Castellanos-Jankiewicz told The Intercept, though these suits were thrown out following a 2016 Supreme Court decision. Most recently Lafarge, a French cement company, pleaded guilty in both the U.S. and France following criminal investigations into its business in ISIS-controlled Syria.

There is essentially no legal precedent as to whether the provision of software to a military committing atrocities makes the software company complicit in those acts. For any court potentially reviewing this, an important legal standard, Castellanos-Jankiewicz said, is whether “Google knew or should have known that its equipment that its software was being either used to commit the atrocities or enabling the commission of the atrocities.”

The Nimbus deal was inked before Hamas attacked Israel on October 7, 2023, igniting a war that has killed tens of thousands of civilians and reduced Gaza to rubble. But that doesn’t mean the company wouldn’t face scrutiny for continuing to provide service. “If the risk of misuse of a technology grows over time, the company needs to react accordingly,” said Andreas Schüller, co-director of the international crimes and accountability program at the European Center for Constitutional and Human Rights. “Ignorance and an omission of any form of reaction to an increasing risk in connection with the use of the product leads to a higher liability risk for the company.”

Though corporations are generally exempt from human rights obligations under international frameworks, Google says it adheres to the United Nations Guiding Principles on Business and Human Rights. The document, while voluntary and not legally binding, lays out an array of practices multinational corporations should follow to avoid culpability in human rights violations.

Among these corporate responsibilities is “assessing actual and potential human rights impacts, integrating and acting upon the findings, tracking responses, and communicating how impacts are addressed.”

The board of directors at Alphabet, Google’s parent entity, recently recommended voting against a shareholder proposal to conduct an independent third-party audit of the processes the company uses “to determine whether customers’ use of products and services for surveillance, censorship, and/or military purposes contributes to human rights harms in conflict-affected and high-risk areas.” The proposal cites, among other risk areas, the Project Nimbus contract. In rejecting the proposal, the board touted its existing human rights oversight processes, and cites the U.N. Guiding Principles and Google’s “AI Principles” as reason no further oversight is necessary. In February, Google amended this latter document to remove prohibitions against weapons and surveillance.

“The UN guiding principles, plain and simple, require companies to conduct due diligence,” said Castellanos-Jankiewicz. “Google acknowledging that they will not be able to conduct these screenings periodically flies against the whole idea of due diligence. It sounds like Google is giving the Israeli military a blank check to basically use their technology for whatever they want.”

The post Google Worried It Couldn’t Control How Israel Uses Project Nimbus, Files Reveal appeared first on The Intercept.

California state Sen. Scott Wiener’s opposition might seem uncharacteristic for someone running a progressive campaign, but it’s consistent with the priorities of two top donors to a super PAC backing his candidacy.

Crypto mogul Chris Larsen and venture capitalist Garry Tan — a pair of wealthy Bay Area tech executives funding a pro-Wiener super PAC called Abundant Future — have been outspoken advocates of stopping the taxes, both of which aim to help fill funding gaps in healthcare and social services after the Trump administration’s recent cuts to Medicaid. Larsen has poured millions of dollars into the fight.

The statewide tax, known as the Billionaire Tax Act, would levy a one-time 5 percent tax on the state’s billionaires’ wealth and assets. The local San Francisco proposition, colloquially known as the Overpaid CEO tax, would tax companies whose CEO makes 100 times more than their median worker, which mostly applies to companies with billionaire CEOs. Both will likely be on the ballot in November, as Wiener also hopes to be.

Larsen, the billionaire co-founder and executive chairman of the blockchain service Ripple Labs and now a mainstay in Bay Area political funding, has donated $100,000 to the PAC backing Wiener — the most of any individual donor — and $700,000 opposing the Overpaid CEO tax, according to federal and San Francisco city records. He’s spent far more fighting the statewide billionaires’ tax, sinking $5 million of his own wealth and another $5 million from Ripple into the Golden State Promise PAC, an anti-tax PAC he founded, per state records. Larsen gave an additional $2.5 million to a separate anti-billionaire tax group, Building a Better California, founded by Google co-founder Sergey Brin and former Google CEO Eric Schmidt. (Brin has reportedly already left the state to avoid the tax.)

Tan, the CEO of startup incubator Y Combinator, has less money to throw around, but he’s made vocal opposition to the tax measures a key part of his brand. He frequently invokes the specter of billionaires and startups fleeing the state and spreads claims that the statewide tax would mean Google’s founders would owe 50 percent of their stocks, which the tax’s backers have dismissed as false. He’s contributed $25,000 to Abundant Future.

Related

She Opposed His Plan for a Blockchain City. Now He’s Bankrolling Her Primary Opponent.

Larsen and Tan likely see their support as “political investments that they expect a return on,” said Jeremy Mack, executive director of Phoenix Project, which tracks corporate spending in San Francisco politics. Wiener owes much of his political strength to the donors who have boosted his housing causes during his state Senate career, including Larsen and Tan. With those backers now animated against the wealth taxes, Mack said that supporting them would be “political suicide” for Wiener.

But Wiener’s opposition to the taxes positions him against the political currents now driving the Democratic Party’s progressive wing. California’s major labor unions, a supermajority of San Francisco’s board of supervisors, and national progressive leaders like Sen. Bernie Sanders, I-Vt., all support the pair of taxes. Even Pelosi, Wiener’s would-be predecessor and a known moderate, is in favor of the local San Francisco tax. SEIU California, one of the state’s largest labor unions, withdrew its endorsement of Wiener in early April over his opposition to the tax measures.

Related

Saikat Chakrabarti’s Plan for the Political Revolution

Both of Wiener’s opponents in the three-way June 2 primary — progressive member of San Francisco’s board of supervisors Connie Chan and Justice Democrats co-founder Saikat Chakrabarti — are in favor of the taxes. Most California voters support the statewide billionaire tax, according to a March poll, including 72 percent of Democratic voters. 

“If you look at who is bankrolling [Wiener], he is doing the bidding of massive corporate interest,” Justin Dolezal, a San Francisco bar owner and co-founder with Small Business Forward, an advocacy group that supports both wealth taxes, told The Intercept. “That’s what he’s looking out for, rather than the average, everyday working San Franciscans.”

Wiener’s campaign did not respond to The Intercept’s request for comment. 

While Wiener in the past has brushed off concerns of corporate backers influencing his policy, saying that he and his wealthiest donors “have agreements and disagreements,” their alignment in opposition against two popular wealth taxes has drawn concern from housing and homelessness advocates, who were already skeptical of Wiener for boosting housing development in the city that they argue favors real estate corporations. The real estate industry was consistently among his top donors during his state Senate elections.

Wiener is a proponent of the “Yes in My Backyard” movement that seeks to address the housing crisis by increasing the housing stock, while opponents criticize it for its emphasis on boosting development rather than redistributing wealth. The movement has morphed over the past several years with the growth of the abundance movement, which is popular among San Francisco’s powerful billionaires and aims to remove regulations and red tape to speed up development.

In addition to being top donors to Abundant Future, Tan and Larsen, along with Yelp CEO Jeremy Stoppleman, have been consistent supporters of Wiener’s YIMBY vision. During his decade in the state Senate, Wiener introduced a series of bills that cut regulations to accelerate housing development across the state, a core tenet of YIMBYism and abundance. Critics on the left dismissed his policies as rewards for corporate commercial real estate developers that failed to meet San Francisco and the state’s housing needs, as well as exacerbating gentrification and displacement of its low-income residents. Opponents instead argue for redistribution of wealth, using the housing that already exists and direct investment in services for low-income people. 

Confronting challenges over his support from wealthy donors during his campaign for Congress, Wiener often refers to his track record of taking on corporations, such as introducing AI regulation bills, one of which drew the ire of some of his tech backers, including Tan. But earlier this year, Wiener and Tan partnered on a failed state bill that would have restricted Big Tech companies from self-preferencing their products over smaller companies. While Wiener touted the legislation as a way to rein in the likes of Apple and Google, Tan’s company, Y Combinator, likely would have benefited because it helps launch new startups.

Tan has also worked to insulate the tech sector from organized labor, accusing the state’s labor leaders of having the goal of “killing the tech golden goose and taking maximum waste into the budget … until CA ceases to work for everyday Californians.” 

Larsen, meanwhile, railed against unions at a San Francisco business event in January, calling on his peers to “start fighting on par with the unions when they propose these absolutely stupid propositions like this crazy CEO tax.” Larsen echoed the message at a separate tech donor gathering Tan hosted months later. 

Larsen did not respond to The Intercept’s request for comment. A spokesperson for Tan told The Intercept to “look at Mr. Tan’s posts on X/Twitter,” where Tan has called the billionaire tax “a destroy tech in California proposition” and the overpaid CEO tax “bad policy wrapped up in anti-billionaire bullshit.”

Wiener’s legislative record reveals an inconsistent history of supporting progressive taxation. In 2018, he opposed a successful local tax on big businesses to fund homelessness services. Two years later, Wiener supported the first iteration of the CEO tax, the first of its kind nationwide, before it was undone in 2024. 

At a candidate forum in January, Wiener said he supported progressive taxes, but he would wait until the Billionaires Tax Act got on the ballot to decide. In April, Wiener said he opposed the local CEO tax, saying he didn’t want to interrupt San Francisco Mayor Daniel Lurie’s economic recovery agenda and that he would pursue similar progressive tax reform in Congress. And last week, after the state billionaire tax’s backers announced they had the necessary signatures to enter it on the ballot, Wiener said he was also against the statewide tax.

Related

Progressive Group Founded by Bernie Sanders Endorses Billionaire for California Governor

“California already has an unstable boom-bust tax system because of the devaluation of property taxes and reliance increasingly on income taxes on wealthy residents,” Wiener told the San Francisco Standard. He said he disagreed with the approach, especially given that it’s a one-time tax.

“It sounds like a person that’s in opposition, but doesn’t want to be seen as Republican,” said Paul Boden, a longtime advocate for people living unhoused. “It’s the neoliberal justification for continuing down the same neoliberal path since Reagan: that doing something that might impact some wealthy people is bad for all of us.” 

Boden, the executive director of the Western Regional Advocacy Project, has long sparred with Wiener on his housing and homelessness policy. In 2016, when Wiener was a San Francisco board supervisor, Boden spoke out against a letter Wiener wrote to the city’s police chief, which had called for a sweep of homeless encampments amid that year’s winter storms. He has criticized Wiener’s housing policies, arguing they prioritize middle-income San Franciscans over the city’s poor.

The results of Larsen and Tan’s ad spending can already be seen on the airwaves in and around San Francisco. Abundant Future has been running ads and sending mailers that paint Chakrabarti, who is advocating to nationalize AI by turning struggling AI companies into public utilities, as a carpetbagger amid his surge in recent polls. Larsen has said that he supports candidates promoting AI regulation, and he plans to spend millions backing Alex Bores, a New York congressional candidate facing heavy oppositional spending from a PAC backed by openAI.

Larsen-funded ads released by his Golden State Promises PAC aired during California’s recent gubernatorial debate, saying the billionaire tax would “backfire and hurt you.”

Supporters of the local and state wealth taxes argue that more revenue is needed to address California’s shortfall due to federal healthcare funding cuts, which is estimated at a $100 billion loss over the next five years. There are more than 200 billionaires who live in the state, according to Forbes data compiled by tax advocates. Most of the revenue from the one-time state tax would go to healthcare, with some set aside for food assistance at schools and other education programs. 

Revenue from San Francisco’s local Overpaid CEO tax — which has been estimated to bring in $250 to $300 million each year — is designed to go to the city’s general fund, with its supporters hoping to invest in healthcare, mental health treatment, and housing support. Larsen and opponents are also funding support for a dueling “poison pill” measure, which would negate the Overpaid CEO tax if approved.

To Mack of the Phoenix Project, this kind of spending is par for the course in politics but should inspire voters to think critically about whom they support.

“The more politicians are in their pockets,” said Mack, referring to wealthy donors, “the less we can expect regular Californian/San Franciscan people’s voices to matter.”

Correction: May 14, 2026, 4:05 p.m. ET
A previous version of this article misstated the first name of a San Francisco bar owner and co-founder with Small Business Forward; he is Justin Dolezal, not Jerome.

The post This California Congressional Hopeful Opposes a Billionaire Tax. So Do His Tech CEO Backers. appeared first on The Intercept.

The student visa form requires applicants to provide the usernames for “each social media platform you have used within the last five years” from a list of 20 specified services, some of them obsolete. This means applicants could find themselves in the awkward position of being required to make public their profiles on the short-form video service Vine, which closed in 2017; the short-lived social media platform Google+, which shut down in 2019; or the dating site Twoo, which ceased operations in 2021.

Most U.S. visa applicants have been required to disclose their profile names on social media accounts since 2019. The Trump administration rolled out new requirements for those seeking student visas under an “expanded screening and vetting” process. The expanded scrutiny applies to F (academic students), M (vocational students), and J (exchange visitor) visa applicants.

Related

Crossing the U.S. Border? Here’s How to Protect Yourself

According to a State Department cable, obtained by the Free Press and Politico, the provided social media accounts will subsequently be checked for “any indications of hostility towards the citizens, culture, government, institutions or founding principles of the United States.”

The DS-160: Online Nonimmigrant Visa Application form doesn’t appear to have been updated to reflect State’s recent guidance, as it doesn’t presently make any mention of the accounts needing to be made public.

“Government social media surveillance invades privacy and chills freedom of speech, and it is prone to errors and misinterpretation without ever having been proven effective at assessing security threats,” warned Sophia Cope, a senior staff attorney at the Electronic Frontier Foundation. She said that by requiring social media accounts be made public, “the U.S. government is endorsing the violation of a fundamental principle of privacy hygiene.”

The online visa application lists a dropdown menu with 20 social media accounts to choose from.

• Ask.fm
• Douban
• Facebook
• Flickr
• Google+
• Instagram
• LinkedIn
• Myspace
• Pinterest
• Qzone (QQ)
• Reddit
• Sina Weibo
• Tencent Weibo
• Tumblr
• Twitter
• Twoo
• Vine
• VKontakte (VK)
• Youku
• YouTube

The list is mishmash of popular social media providers, regional services (predominantly those used in China), and a bevy of outdated and defunct platforms, such as Myspace, which has been a digital ghost town for years.

A quarter of the sites listed no longer exist at all, with some already being defunct when the visa application form first started requiring the disclosure of social media usernames in 2019. That includes Ask.fm, a Latvian service where users could ask questions that closed last year, and Tencent Weibo, a Chinese microblogging service that shut down in 2020.

Among the included services are Douban, Qzone, Sina Weibo, and Youku — all active Chinese social network sites. Despite listing five different Chinese social media sites, the form leaves off Tencent’s WeChat, China’s most popular social media app.

VKontakte is the only Russian social media service appearing on the list. No other popular regional social media sites are included.

Other modern social media platforms, such as TikTok or Trump’s own Truth Social, are missing from the list as well, though the visa form does allow applicants to specify additional accounts.

Asked for comment on how this list of social media platforms was compiled, or whether there are plans to update the online form, a State Department spokesperson provided a statement summarizing the new guidance and said that “the Trump Administration is focused on protecting our nation and our citizens by upholding the highest standards of national security and public safety through our visa process.”

Albert Fox Cahn, founder and executive director of the Surveillance Technology Oversight Project, described the policy as “antithetical to everything our First Amendment should protect,” pointing out that “not only will these shortsighted efforts fail to protect the public, they’ll put countless students at risk. Now those who wanted to study in the U.S. to flee authoritarian governments abroad will have to make their social media public to those same governments to study here.”

The post Want a Student Visa? The U.S. Government Needs Your Vine Account. appeared first on The Intercept.

U.S. Customs and Border Protection is planning to modernize older video surveillance towers in Arizona that provide the agency an unblinking view of the border. A key part of the effort is adding machine-learning capabilities to CBP cameras, allowing the agency to automatically detect humans and vehicles as they approach the border without continuous monitoring by humans. CBP is purchasing computer vision powers from two vendors, IBM and Equitus. Google, the documents show, will play a critical role stitching those services together by operating a central repository for video surveillance data.

Related

The U.S. Border Patrol and an Israeli Military Contractor Are Putting a Native American Reservation Under “Persistent Surveillance”

The work is focused on older towers purchased from Israeli military defense contractor Elbit. In all, the document notes “50 towers with up to 100 cameras across 6 sites in the Tucson Sector” will be upgraded with machine learning capabilities.

IBM will provide its Maximo Visual Inspection software, a tool the company generally markets for industrial quality control inspections — not tracking humans. Equitus is offering its Video Sentinel, a more traditional video surveillance analytics program explicitly marketed for border surveillance that, according to a promotional YouTube video, recently taken offline, featuring a company executive, can detect “people walking caravan style … some of them are carrying backpacks and being identified as mules.”

“Within 60 days from the start of the project, real life video from the southern border is available to train and create AI/ML models to be used by the Equitus Video Sentinel.”

Tying together these machine learning surveillance tools is Google, which the document reveals is supplying CBP with a cloud computing platform known as MAGE: the ModulAr Google Cloud Platform Environment. Based on the document, Google is providing a hub for video surveillance data and will directly host the Equitus AI analysis tool. It appears every camera in CBP’s Tucson Sector will pipe data into Google servers: “This project will focus initially on 100 simultaneous video streams from the data source for processing,” the document reads, and “the resulting metadata and keyframes will be sent to CBP’s Google Cloud.”

The diagram also notes that one of Google’s chief rivals, Amazon Web Services, provides CBP with unspecified cloud computing services.

During President Trump’s first term, border surveillance and immigration enforcement work carried a stigma in the tech sector it has in part shed today.

Related

Google AI Tech Will Be Used for Virtual Border Wall, CBP Contract Shows

In 2020, The Intercept revealed a document produced by the CBP Innovation Team, known as INVNT, that stated Google Cloud services would be used in conjunction with AI-augmented surveillance towers manufactured by defense contractor Anduril: “Google Cloud Platform (GCP) will be utilized for doing innovation projects for C1’s INVNT team like next generation IoT, NLP (Natural Language Processing), Language Translation and Andril [sic] image camera and any other future looking project for CBP. The GCP has unique product features which will help to execute on the mission needs.” (A CBP spokesperson confirmed to The Intercept that “Andril” was a misspelling of Anduril.)

After the Anduril work came to light, Google’s cloud computing chief Thomas quickly attempted damage control, directly contradicting the Department of Homeland Security and telling concerned employees that the company was not involved in immigration enforcement on the Mexican border, CNBC reported at the time. “We have spoken directly with Customs and Border Patrol and they have confirmed that they are not testing our products for those purposes,” Kurian added.

If this was true then, it’s certainly not now; references to Google services appear repeatedly throughout the tower modernization project document. A technical diagram showing how video data flows between various CBP servers shows Google’s MAGE literally in the middle.

Customs and Border Protection did not respond to a request for comment about its use of Google Cloud.

Google did not respond to specific questions about the project, nor address Kurian’s prior denial.

In a statement provided to The Intercept, Google Public Sector executive Jim Kelly attempted to distance the company slightly from the border surveillance work. “CBP has been public about how it has a multicloud strategy and has used Google Cloud for work like translation,” Kelly wrote. “In this case, Google Cloud is not on the contract. That said, customers or partners can purchase Google Cloud’s off-the-shelf compute, storage, and networking products for their own use, much like they might use a mobile network or run their own computer hardware.”

Kelly’s statement indicates the government is acquiring Google Cloud services through a reseller, as is common in federal procurement. But Kelly’s comparison of Google Cloud technology to buying off-the-shelf computer hardware is misleading. Even if it’s supplied through a subcontractor or reseller, CBP’s use of Google’s service still requires a constant and ongoing connection to the company’s cloud infrastructure. Were Google still serious about “not working on any projects associated with immigration enforcement at the southern border,” as Kurian claimed in 2020, it would be trivial to prevent CBP from using Google Cloud.

Industry advocates and immigration hard-liners have long touted the “virtual wall” initiative, which substitutes iron and concrete barriers and Border Patrol agents for a 2,000 mile array of sensors, cameras, and computers. But critics say advanced technology is no substitute for policy reforms.

“On top of the wasted tax dollars, border communities end up paying the price with their privacy, as demonstrated by the recent findings by the Government Accountability Office that CBP had failed to implement six out of six key privacy policy requirements,” Dave Maass, director of investigations at the Electronic Frontier Foundation, told The Intercept, referring to the tower program’s dismal privacy protections record. “For more than two decades, surveillance towers at the border have proven to be a boondoggle, and adding AI isn’t going to make it any less of a boondoggle — it will just be an AI-powered boondoggle.”

The post Google Is Helping the Trump Administration Deploy AI Along the Mexican Border appeared first on The Intercept.

YouTube surreptitiously deleted all these videos in early October by wiping the accounts that posted them from its website, along with their channels’ archives. The accounts belonged to three prominent Palestinian human rights groups: Al-Haq, Al Mezan Center for Human Rights, and the Palestinian Centre for Human Rights.

The move came in response to a U.S. government campaign to stifle accountability for alleged Israeli war crimes against Palestinians in Gaza and the West Bank.

The Palestinian groups’ YouTube channels hosted hours of footage documenting and highlighting alleged Israeli government violations of international law in both Gaza and the West Bank, including the killing of Palestinian civilians.

“I’m pretty shocked that YouTube is showing such a little backbone,” said Sarah Leah Whitson, executive director of Democracy for the Arab World Now. “It’s really hard to imagine any serious argument that sharing information from these Palestinian human rights organizations would somehow violate sanctions. Succumbing to this arbitrary designation of these Palestinian organizations, to now censor them, is disappointing and pretty surprising.”

After the International Criminal Court issued arrest warrants and charged Israeli Prime Minister Benjamin Netanyahu and former Israeli Defense Secretary Yoav Gallant with war crimes in Gaza, the Trump administration escalated its defense of Israel’s actions by sanctioning ICC officials and targeting people and organizations that work with the court.

“YouTube is furthering the Trump administration’s agenda to remove evidence of human rights violations and war crimes.”

“It is outrageous that YouTube is furthering the Trump administration’s agenda to remove evidence of human rights violations and war crimes from public view,” said Katherine Gallagher, a senior staff attorney at the Center for Constitutional Rights. “Congress did not intend to allow the president to cut off the flow of information to the American public and the world — instead, information, including documents and videos, are specifically exempted under the statute that the president cited as his authority for issuing the ICC sanctions.”

“Alarming Setback”

YouTube, which is owned by Google, confirmed to The Intercept that it deleted the groups’ accounts as a direct result of State Department sanctions against the group after a review. The Trump administration leveled the sanctions against the organizations in September over their work with the International Criminal Court in cases charging Israeli officials of war crimes.

“Google is committed to compliance with applicable sanctions and trade compliance laws,” YouTube spokesperson Boot Bullwinkle said in a statement.

According to Google’s Sanctions Compliance publisher policy, “Google publisher products are not eligible for any entities or individuals that are restricted under applicable trade sanctions and export compliance laws.”

Al Mezan, a human rights organization in Gaza, told The Intercept that its YouTube channel was abruptly terminated this year on October 7 without prior notification.

“Terminating the channel deprives us from reaching what we aspire to convey our message to, and fulfill our mission,” a spokesperson for the group said, “and prevents us from achieving our goals and limits our ability to reach the audience we aspire to share our message with.”

The West Bank-based Al-Haq’s channel was deleted on October 3, a spokesperson for the group said, with a message from YouTube that its “content violates our guidelines.”

Related

Palestinian Rights Groups That Document Israeli Abuses Labeled “Terrorists” by Israel

“YouTube’s removal of a human rights organisation’s platform, carried out without prior warning, represents a serious failure of principle and an alarming setback for human rights and freedom of expression,” the Al-Haq spokesperson said in a statement. “The U.S. Sanctions are being used to cripple accountability work on Palestine and silence Palestinian voices and victims, and this has a ripple effect on such platforms also acting under such measures to further silence Palestinian voices.”

The Palestinian Center for Human Rights, which the U.N. describes as the oldest human rights organization in Gaza, said in a statement that YouTube’s move “protects perpetrators from accountability.”

“YouTube’s decision to close PCHR’s account is basically one of many consequences that we as an organisation have faced since the decision of the US government to sanction our organisations for our legitimate work,” said Basel al-Sourani, an international advocacy officer and legal advisor for the group. “YouTube said that we were not following their policy on Community Guidelines, when all our work was basically presenting factual and evidence-based reporting on the crimes committed against the Palestinian people especially since the start of the ongoing genocide on 7 October.”

“By doing this, YouTube is being complicit in silencing the voices of Palestinian victims,” al-Sourani added.

Looking Outside the U.S.

The three human rights groups’ account terminations cumulatively amount to the erasure of more than 700 videos, according to an Intercept tally.

The deleted videos range in scope from investigations, such as an analysis of the Israeli killing of American journalist Shireen Abu Akleh, to testimonies of Palestinians tortured by Israeli forces and documentaries like “The Beach,” about children playing on a beach who were killed by an Israeli strike.

Some videos are still available through copies saved on the Internet Archive’s Wayback Machine or on alternate platforms, such as Facebook and Vimeo. The wiping only affected the group’s official channels; videos which were produced by the nonprofits but hosted on alternate YouTube channels remain active. No cumulative index of videos deleted by YouTube is available, however, and many appear to not be available elsewhere online.

Videos posted elsewhere online, the groups fear, could soon be targeted for deletion because many of the platforms hosting them are also U.S.-based services. The ICC itself began exploring using service providers outside the U.S.

Al-Haq said it would also be looking for alternatives outside of U.S. companies to host their work.

YouTube isn’t the only U.S. tech company blocking Palestinian rights groups from using its services. The Al-Haq spokesperson said Mailchimp, the mailing list service, also deleted the group’s account in September. (Mailchimp and its parent company, Intuit, did not immediately respond to a request for comment.)

Caving to Trump’s Demand

Both the U.S. and Israeli governments have long shielded themselves from the ICC and accountability for their alleged war crimes. Neither country is party to the Rome Statute, the international treaty that established the court.

In November 2024, the ICC prosecutors issued arrest warrants for Netanyahu and Gallant, charging the leaders with intentionally starving civilians by blocking aid from entering into Gaza. Both the Biden and Trump administrations rejected the legitimacy of the warrants.

Related

Trump Sanctions Palestinian Human Rights Groups for Doing Their Job. Anybody Could Be Next.

Since his reelection, Trump has taken a more aggressive posture against accountability for Israel. In the early days of his second term, Trump renewed sanctions against the ICC and issued new, more severe measures against court officials and anyone accused of aiding their efforts. In September, in a new order, he specifically sanctioned the three Palestinian groups.

The U.S. moves followed Israel’s own designation of Al-Haq as a “terrorist organization” in 2021 and an online smear campaign by pro-Israeli activists attempting to link Palestinian Centre for Human Rights with militant groups.

The sanctions freeze the organizations’ assets in the U.S. and bar sanctioned individuals from traveling to the country. Federal judges have already issued preliminary injunctions in two cases in favor of plaintiffs who argued the sanctions had violated their First Amendment rights.

“The Trump administration is focused on contributing to the censorship of information about Israeli atrocities in Palestine and the sanctions against these organizations is very deliberately designed to make association with these organizations frightening to Americans who will be concerned about material support laws,” said Whitson, of DAWN, which joined a coalition of groups in September to demand the Trump administration drop its sanctions.

Like many tech firms, YouTube has shown a ready willingness to comply with demands from both the Trump administration and Israel. YouTube coordinated with a campaign organized by Israeli tech workers to remove social media content deemed critical of Israel. At home, Google, YouTube’s parent company, secretly handed over personal Gmail account information to U.S. Immigration and Customs Enforcement in an effort to detain a pro-Palestinian student organizer.

Even before Israel’s genocidal campaign in Gaza, YouTube had been accused of unevenly applying its community guidelines to censor Palestinian voices while withholding similar scrutiny from pro-Israeli content. Such trends continued during the war, according to a Wired report.

Earlier this year, YouTube shut down the official account of the Addameer Prisoner Support and Human Rights Association. The move came after pressure from UK Lawyers for Israel, which wrote to YouTube to point out that the organization had been sanctioned by the State Department.

Whitson warned that YouTube’s capitulation could set a precedent, pushing other tech companies to bend to censorship.

“They are basically allowing the Trump administration to dictate what information they share with the global audience,” she said. “It’s not going to end with Palestine.”

The post YouTube Quietly Erased More Than 700 Videos Documenting Israeli Human Rights Violations appeared first on The Intercept.

Many of the assembled companies are not household names in the United States, but several multinational firms — like Nokia, Dell, and Canon — were present at July 10 event.

The mission they had gathered to support was clear. Onstage, a brigadier general with the Israeli military gave a presentation that connected the Nakba, the 1967 Arab–Israeli War, the 2006 invasion of Lebanon, the current war on Gaza, and more wars in the decades to come. His call to action splashed across the big screen: “Each generation and its own turn — this is our watch!”

One company, however, was conspicuously absent: Google.

For the last two years, Google had been a marquee sponsor of IT For IDF — the company is a natural partner for the event, given Google Cloud’s foundational role in Project Nimbus, a $1.2 billion contract aimed at modernizing cloud computing operations across Israeli government that it shares with Amazon.

This year was supposed to be no different and, until just days before the conference started, a Google Cloud logo was displayed alongside other sponsors on the IT For IDF website. Then Google abruptly vanished from the site without explanation.

When asked by a news outlet about the logo’s disappearance, the conference organizers claimed they weren’t aware Google had been on their website and suggested its inclusion was an error. “It’s possible that we used their logo by mistake but they are not a sponsor,” a spokesperson told 404 Media, “as far as I know.”

Google’s own corporate schedule, reviewed by The Intercept, seems to contradict this statement. The document includes upcoming Google events in Israel, and IT For IDF 2024 is on the list. On this internal schedule, Google is explicitly labeled as a co-sponsor of the conference in partnership with CloudEx, an Israeli cloud computing consultancy.

CloudEx CEO Ariel Munafo moonlights as an adviser to the IDF’s Center of Computers and Information Systems, known as Mamram, where he is helping other IDF units build out their cloud computing operations, according to his LinkedIn profile.

Google did not respond to multiple requests for comment.

The apparent about-face on Google’s sponsorship of IT For IDF is just one recent example of companies seeking to distance themselves from Israel’s brutal assault on the Gaza Strip, which has killed nearly 40,000 Palestinians. Though much business has gone on as usual amid the destruction, some companies have, at various times, and to various extents, shifted their business plans in Israel.

Google continues to work with the IDF, as it has for years on the Nimbus contract. The company’s odd vanishing act from a conference focused on a lucrative customer relationship stands as one of the most high-profile examples of what appears to be PR anxiety.

The tech giant has shown some squeamishness over some of Nimbus’s objectives in the past. The project has drawn international criticism and prompted a dissent campaign among Google employees, over 50 of whom were fired in April for protesting the contract. The Israeli government emphasizes Nimbus’s military dimensions, but Google has persistently tried to downplay or outright deny that its contract for Israel includes military work.

Official event photos posted to a public album and reviewed by The Intercept suggest that Google’s connection to the IDF networking event was literally whitewashed before it began. Official photos of a step-and-repeat backdrop for the conference contains all the event’s original sponsors’ logos, sans Google — and includes one square that appears to have been hastily painted over. Neither Google nor People and Computers, the conference’s organizer, responded when asked whose logo was underneath the paint.

The logo of Cisco, which claimed to 404 Media that it was “not a sponsor of this conference,” remained on display.

Even without a presence at the conference, Google loomed over an event dedicated to the prosecution of a war from which it has struggled to distance itself.

One of the event’s big draws was a presentation by Col. Racheli Dembinski of Mamram, on the Israeli military’s use of cloud platforms during the war in Gaza, during which she highlighted the role of Google Cloud, according to Israeli press.

A later talk presented by the “technology lead for Google Cloud Platform and CloudEx” noted that CloudEx’s partnership with Google entailed “working closely on several production cloud-hosted projects” with the Ministry of Defense.

Commit, another cloud computing reseller that was recently named “2024 Google Cloud Sales Partner of the Year for Israel” for its work implementing Project Nimbus, took to the stage to hawk its battlefield management software.

Related

Documents Reveal Advanced AI Tools Google Is Selling to Israel

Google’s insistence that Project Nimbus is peaceful in nature is at odds with the public record. Nimbus training materials published by The Intercept in 2022 cited the Ministry of Defense as a customer. Recent reporting by Wired detailed the project’s connection to the IDF since its inception. When it was announced in 2021, Project Nimbus was touted by the Israeli Finance Ministry as serving the “defense establishment.”

In May, journalist Jack Poulson reported on Google’s long-running collaboration with IDF information technology units like Mamram, noting an intensification in such work since October 7.

Despite all this, Google has routinely refused to discuss the Nimbus work or its humanitarian implications with the press. The company generally responding to inquiries about the project or criticism of its military nature with a boilerplate statement that Nimbus is “not directed at highly sensitive or classified military workloads relevant to weapons or intelligence services.”

The Intercept revealed in May that Nimbus requires two prominent Israeli weapons manufacturers, Israel Aerospace Industries and Rafael Advanced Defense Systems, to use Google and Amazon cloud computing services in their work.

The post Google Planned to Sponsor IDF Conference That Now Denies Google Was Sponsor appeared first on The Intercept.

Days before the federal government falsely claimed cellphone-brandishing nurse Alex Pretti was a terrorist plotting a “massacre,” a jury in Chicago acquitted Juan Espinoza Martinez on bogus charges of a murder-for-hire plot against then-Border Patrol commander Greg Bovino. A recently unsealed court transcript shows the government used that case to bolster its claims about the dangers of “doxing” Immigration and Customs Enforcement officials. That pretext was used to convince a judge to obscure an ICE agent’s face during a public court proceeding when his name, face, employment, and location were publicly listed on his LinkedIn page.  

As with its baseless claims about Pretti, the government presented no evidence supporting its proclamations that Martinez, a union carpenter, was a higher-up in the Latin Kings gang with the ability or intent to put out hits on Bovino or other immigration agents. The case against him hinged on ambiguous Snapchat messages that Martinez’s attorney called “neighborhood gossip.” But the Department of Homeland Security brought its allegations to the public long before it could be tested in court, repeating claims of bounties up to $50,000. 

The transcript from a federal court in Chicago, which was recently released pursuant to a motion filed by law firm Mandell PC on behalf of local media outlets, shows how far the hysteria has gone. During an October 20, 2025, hearing in a case challenging immigration enforcement tactics, government lawyers asked for a private conference with Judge Sara Ellis to request the courtroom sketch artist not draw ICE Deputy Field Office Director Shawn Byers.

Government attorneys claimed that, in light of the alleged “bounties” on the heads of ICE agents, Byers had taken extensive precautions to disconnect his identity from his image online to protect himself. When the judge asked for details on the bounties, Department of Justice attorney Samuel Holt responded, “I don’t have all the details. My understanding is that I — I think it was a gang bounty.”

The judge cleared the courtroom and called Byers in to provide the details about the “threat.” Byers first claimed there was a $50,000 “bounty issued by the cartels on me,” along with $10,000 “for all my family members.” He also said the “credible threat” was out against “all senior ICE officials here in Chicago,” where Byers said he was the most senior ICE agent on the ground. Asked when he learned about the bounty, Byer said “It’s been about a week or so I believe.” Martinez’s arrest was announced two weeks earlier, on October 6; no other bounties were publicly reported in the interim. When the judge asked whether these threats were “directed specifically” at him, Byers seemed to walk his claims back, replying, “Well, all senior ICE officials. So it’s not just me.”

Byers also said he’d taken action to “limit social media exposure” and “reduce the footprint” to avoid his face being connected with his name and that even his appearance in court required “additional precautions.”

“You know, my name is out there. I’ve been doxed as — as recently as over the weekend,” Byers told the judge, according to the court transcript. “So my name is out there, but my name has not been connected to my face yet, so that’s what I’m trying to prevent from happening.”

Despite objections from opposing counsel that court proceedings (and courtroom sketches) should be public, the judge ordered the sketch artist to blur Byers’s facial features, concealing his identity. Ellis’s compromise, while likely intended as a good-faith effort to balance safety and transparency, nonetheless validated the notion that immigration agents operate under extreme risk, justifying extraordinary protective measures by our legal system. It also effectively brought the masks immigration agents wear on the street into the courtroom.

Then, while Byers and other witnesses testified, someone apparently Googled his name and informed the judge that a simple search turned up his LinkedIn profile, complete with his photo, his exact job title, and his location in Chicago. 

The judge called the parties back into closed session (it’s unclear why, given that the false reason for the earlier private sidebar had been exposed). 

“I got to say, you know, I feel slightly foolish in trying to protect Mr. Byers when, you know, a simple Google search pulls up his name and his picture,” she said, according to the transcript. She also encouraged the attorney to advise the ICE deputy director that his name and photograph were readily available online. “If I could find his picture in two seconds with his name, it just looks a little silly to be asking the courtroom sketch artist to blur his features.” Being recognized is “the cost of being a public servant,” she continued. 

The judge also said moving forward, she would “just be more hesitant to kind of obscure somebody’s identity,” but did not say she’d be entering any actual sanctions for the half-baked rationale used to convince her to censor the public record. 

After some back and forth with the DOJ attorneys about whether Byers’s LinkedIn profile contained his actual picture, Ellis confirmed the profile for “Shawn B.” did when viewed by someone logged into LinkedIn. (A LinkedIn search for “Shawn Byers ICE” brings up just one profile for a Shawn B., who is listed as currently working as Deputy Field Office Director for ICE in Chicago. It also notes he is a 22-year veteran of the department and contains reposts about ICE removals in Chicago and a hiring notice for GEO Group, the for-profit prison conglomerate contracted with ICE, but no longer contains any profile picture.)

Related

Trump Calls His Enemies Terrorists. Does That Mean He Can Just Kill Them?

Since Byers’s manufactured emergency obviously wasn’t based on real concerns for his safety, what was the point of the whole sideshow? It was likely intended to feed the narrative that immigration agents face such grave threats that identifying them — in addition to filming their operations, following them to do so, tracking and communicating about their locations and other clearly constitutionally protected conduct — needs to be restrained. It’s the same fiction that primes segments of the American public to be receptive to claims that people like Pretti and Renee Good were threatening officers’ lives to justify their killings.

In January, Homeland Security Secretary Kristi Noem scolded CBS News’ “Face the Nation” host Margaret Brennan for naming Jonathan Ross, the immigration agent who shot and killed Good in Minneapolis. She accused Brennan of “continu[ing] to dox law enforcement,” despite acknowledging that Ross’s name was already very public, citing unspecified attacks against his family. It’s far from the first time Noem and others have claimed that naming or videotaping law enforcement officers is improper, illegal, or even intended to foment violence.

These efforts to chill the work of reporters and ICE watchers have spread beyond immigration enforcement, as we saw from last month’s subpoena by the House Oversight Committee of journalist Seth Harp, which was accompanied by a criminal referral to the Department of Justice by Rep. Anna Paulina Luna, a Republican from Florida. Harp was also accused of “doxing” for naming a Delta Force commander involved in the abduction of Venezuelan President Nicolás Maduro, an allegation backed up by unsubstantiated claims that the commander’s life was at risk.

The notion that naming public officials at the center of major news stories, who very often conceal their identities while carrying out unprecedented law enforcement operations on the streets of our cities, or that simply drawing their faces for the court record is “doxing” or otherwise improper, is a complete Trump administration fabrication. Still, the government is repeating it often enough that it’s warping the public’s perception of journalism. The Byers ordeal is an unusually clear example of the current playbook being used to shield administration officials and their foot soldiers from accountability under the guise of protecting public officials’ safety. 

The next time this happens in court, the judge needs to demand specifics, with evidence, about whatever nebulous alleged plots or threats the government is pushing to justify secrecy. With comprehensive studies demonstrating their constant misrepresentations, nothing government lawyers say can be taken at face value. And when it happens outside the courthouse, the media needs to be similarly skeptical and not take the “threats” narrative at face value from an administration with a long, proven track record of misleading the public for its own political ends. 

Judges also need to impose significant sanctions on lawyers and witnesses who mislead them, make them pawns in the administration’s anti-transparency objectives, and waste their time. Gently reprimanding them in private doesn’t cut it, especially when these false, alarmist narratives used in court are then being used to justify ICE killings to the public.

The post Judge Censored an ICE Agent’s Face Over “Threats.” His Info Was a Google Search Away. appeared first on The Intercept.

In partnership with

Rita Murad, a 21-year-old Palestinian citizen of Israel and student at the Technion Israel Institute of Technology, was arrested by Israeli authorities in November 2023 after sharing three Instagram stories on the morning of October 7. The images included a picture of a bulldozer breaking through the border fence in Gaza and a quote: “Do you support decolonization as an abstract academic theory? Or as a tangible event?” She was suspended from university and faced up to five years in prison.

In recent years, Israeli security officials have boasted of a “ChatGPT-like” arsenal used to monitor social media users for supporting or inciting terrorism. It was released in full force after Hamas’s bloody attack on October 7. Right-wing activists and politicians instructed police forces to arrest hundreds of Palestinians within Israel and east Jerusalem for social media-related offenses. Many had engaged in relatively low-level political speech, like posting verses from the Quran on WhatsApp or sharing images from Gaza on their Instagram stories.

When the New York Times covered Murad’s saga last year, the journalist Jesse Baron wrote that, in the U.S., “There is certainly no way to charge people with a crime for their reaction to a terrorist attack. In Israel, the situation is completely different.”

Soon, that may no longer be the case.

Hundreds of students with various legal statuses have been threatened with deportation on similar grounds in the U.S. this year. Recent high-profile cases have targeted those associated with student-led dissent against the Israeli military’s policies in Gaza. There is Mahmoud Khalil, a green card holder married to a U.S. citizen, taken from his Columbia University residence and sent to a detention center in Louisiana. There is Rümeysa Öztürk, a Turkish doctoral student at Tufts disappeared from the streets of Somerville, Massachusetts, by plainclothes officers allegedly for co-authoring an op-ed calling on university administrators to heed student protesters’ demands. And there is Mohsen Mahdawi, a Columbia philosophy student arrested by ICE agents outside the U.S. Citizenship and Immigration Services office where he was scheduled for his naturalization interview.

In some instances, the State Department has relied on informants, blacklists, and technology as simple as a screenshot. But the U.S. is in the process of activating a suite of algorithmic surveillance tools Israeli authorities have also used to monitor and criminalize online speech.

In March, Secretary of State Marco Rubio announced the State Department was launching an AI-powered “Catch and Revoke” initiative to accelerate the cancellation of student visas. Algorithms would collect data from social media profiles, news outlets, and doxing sites to enforce the January 20 executive order targeting foreign nationals who threaten to “overthrow or replace the culture on which our constitutional Republic stands.” The arsenal was built in concert with American tech companies over the past two decades and already deployed, in part, within the U.S. immigration system.

Rubio’s “Catch and Revoke” initiative emerges from long-standing collaborations between tech companies and increasingly right-wing governments eager for their wares. The AI industry’s business model hinges on unfettered access to troves of data, which makes less-then-democratic contexts, where state surveillance is unconstrained by judicial, legislative, or public oversight, particularly lucrative proving grounds for new products. The effects of these technologies have been most punitive on the borders of the U.S. or the European Union, like migrant detention centers in Texas or Greece. But now the inevitable is happening: They are becoming popular domestic policing tools.

Israel was one early test site. As Israeli authorities expanded their surveillance powers to clamp down on rising rates of Palestinian terrorism in the early 2010s, U.S. technology firms flocked to the region. In exchange for first digital and then automated surveillance systems, Israel’s security apparatus offered CEOs troves of the information economy’s most prized commodity: data. IBM and Microsoft provided software used to monitor West Bank border crossings. Palantir offered predictive policing algorithms to Israeli security forces. Amazon and Google would sign over cloud computing infrastructure and AI systems. The result was a surveillance and policing dragnet that could entangle innocent people alongside those who posed credible security threats. Increasingly, right-wing ruling coalitions allowed it to operate with less and less restraint.

With time and in partnership with many of the same companies, the U.S. security state built its own surveillance capacities to scale.

Not long ago, Silicon Valley preached a mantra of globalization and integration. It was antithetical to the far-right’s nationalistic agenda, but it was good for business in an economy that hinged on the skilled and unskilled labor of foreigners. So when Trump signed an executive order banning immigration from five Muslim countries and subjecting those approved for visas to extra screening in January 2017, tech executives and their employees dissented.

Google co-founder Sergey Brin, an immigrant from the Soviet Union, joined demonstrations at the San Francisco airport to protest Trump’s travel ban. Mark Zuckerberg cited his grandparents, Jewish refugees from Poland, as grounds for his opposition to the policy. Sam Altman also called on industry leaders to take a stand. “The precedent of invalidating already-issued visas and green cards should be extremely troubling for immigrants of any country,” he wrote on his personal blog. “We must object, or our inaction will send a message that the administration can continue to take away our rights.”

Many tech workers spent the first Trump presidency protesting these more sinister entailments of a data-driven economy. Over the following year, Microsoft, Google, and Amazon employees would stage walkouts and circulate petitions demanding an end to contracts with the national security state. The pressure yielded image restoration campaigns. Google dropped a bid for a $10 million Defense Department contract. Microsoft promised their software and services would not be used to separate families at the border.

Related

Trump’s Election Is Also a Win for Tech’s Right-Wing “Warrior Class”

But the so-called tech resistance belied an inconvenient truth. Silicon Valley firms supplied the software and computing infrastructure that enabled Trump’s policies. Companies like Babel and Palantir entered into contracts with ICE in 2015, becoming the bread and butter of ICE’s surveillance capacities by mining personal data from thousands of sources for government authorities, converting it into searchable databases, and mapping connections between individuals and organizations. By 2017, conglomerates like Amazon, Microsoft, and Google were becoming essential too, signing over the cloud services to host mounds of citizens’ and residents’ personal information.

Even as some firms pledged to steer clear of contracts with the U.S. security state, they continued working abroad, and especially in Israel and Palestine. Investigative reporting over the last year has brought more recent exchanges to light. Deals between U.S. companies and the Israeli military ramped up after October 7, according to leaked documents from Google and Microsoft. Intelligence agencies relied on Microsoft Azure and Amazon Web Services to host surveillance data and used Google’s Gemini and OpenAI’s ChatGPT to cull through and operationalize much of it, often playing direct roles in operations — from arrest raids to airstrikes — across the region.

These contracts gave U.S. technology conglomerates the chance to refine military and homeland security systems abroad until Trump’s reelection signaled they could do so with little pushback at home. OpenAI changed its terms of use last year to allow militaries and security forces to deploy their systems for “national security purposes.” Google did the same this February, removing language saying it wouldn’t use its AI for weapons and surveillance from its “public ethos policy.” Meta also announced U.S. contractors could use its AI models for “national security” purposes.

Technology firms are committed to churning out high-risk products at a rapid pace. Which is why privacy experts say their products can turbocharge the U.S. surveillance state at a time when constitutional protections are eroding.

“It’s going to give the government the impression that certain forms of surveillance are now worth deploying when before they would have been too resource intensive,” Ben Wizner, director of the ACLU’s Speech, Privacy, and Technology Project, offered over the phone last week. “Now that you have large language models, you know, the government may say why not store thousands of hours of conversations just to run an AI tool through them and decide who you don’t want in your country.” 

Related

Google Is Helping the Trump Administration Deploy AI Along the Mexican Border

The parts are all in place. According to recent reports, Palantir is building ICE an “immigrationOS” that can generate reports on immigrants and visa holders — including what they look like, where they live, and where they travel — and monitor their location in real time. ICE will use the database combined with a trove of other AI tools to surveil immigrants’ social media accounts, and to track down and detain “antisemites” and “terrorists,” according to a recent announcement by the State Department. “We need to get better at treating this like a business,” Acting ICE Director Todd Lyons said in a speech at the 2025 Border Security Expo in Phoenix earlier this month, “like [Amazon] Prime, but with human beings.”

 

It is important to remember that many of the propriety technologies private companies are offering the U.S. surveillance state are flawed. Content moderation algorithms deployed by Meta often flag innocuous content as incendiary, especially Arabic language posts. OpenAI’s large language model are notorious for generating hallucinatory statements and mistranslating phrases from foreign languages into English. Stories of error abound in recent raids and arrests, from ICE officials mistaking Mahmoud Khalil for a student visa holder to citizens, lawful residents, and tourists with no criminal record being rounded up and deported.

But where AI falters technically, it delivers ideologically. We see this in Israel and Palestine, as well as other contexts marked by relatively unchecked government surveillance. The algorithms embraced by Israel’s security forces remain rudimentary. But officials have used them to justify increasingly draconian policies. The Haifa-based human rights organization Adalah says there are hundreds of Palestinians with no criminal record or affiliation with militant groups held behind bars because right-wing activists and politicians instructed police forces to search their phones and social media pages and label what they said, shared, or liked online as “incitement to terrorism” or “support of terrorism.”

Now we hear similar stories in American cities, where First Amendment protections and due process are disintegrating. The effects were nicely distilled by Ranjani Srinivasan, an Indian Ph.D. student at Columbia who self-deported after ICE officials showed up at her door and cancelled her legal status. From refuge in Canada, she told the New York Times she was fearful of the U.S. expanded algorithmic arsenal. “I’m fearful that even the most low-level political speech or just doing what we all do — like shout into the abyss that is social media — can turn into this dystopian nightmare,” Srinivasan said, “where somebody is calling you a terrorist sympathizer and making you, literally, fear for your life and your safety.”

It is frightening to think that all this happened in Trump’s first 100 days in office. But corporate CEOs have been working with militaries and security agencies to sediment this status quo for years now. The visible human cost of these exchanges may spawn the opposition needed to head off more repression. But for now, the groundwork is laid for the U.S. surveillance state to finally operate at scale.

The post U.S. Companies Honed Their Surveillance Tech in Israel. Now It’s Coming Home. appeared first on The Intercept.

Related

Documents Reveal Advanced AI Tools Google Is Selling to Israel

The $1.2 billion joint contract split between Google and Amazon provides the Israeli government, including its military, with access to state-of-the-art cloud computing and artificial intelligence tools. This has made Project Nimbus a consistent source of protest inside and outside Google, even before Israel’s war on Gaza.

While Amazon has largely remained silent in the face of employee activism and outside scrutiny, Google routinely downplays or denies the military reach of Project Nimbus — despite the Israeli Finance Ministry’s 2021 announcement that the deal would service the country’s “defense establishment.”

Google has also sought to reassure those concerned by its relationship with a government whose leadership is being investigated by the International Criminal Court for crimes against humanity by claiming Nimbus is constrained by the company’s general rules and regulations.

“We have been very clear that the Nimbus contract is for workloads running on our commercial cloud by Israeli government ministries, who agree to comply with our Terms of Service and Acceptable Use Policy,” a Google spokesperson told Wired in July and repeated verbatim to Time magazine in August, linking both times to the public-facing copies of each document.

Google Cloud’s terms of service prohibit, among other things, uses that “violate, or encourage the violation of, the legal rights of others,” any “invasive” purpose, or anything “that can cause death, serious harm, or injury to individuals or groups of individuals.”

But the premise that Google policy dictates how Nimbus is used is called into question by a previously undisclosed email from a company lawyer. On December 10, 2020, before the tech giant won the contract, Google lawyer Edward du Boulay wrote to company executives with exciting news: “Google Cloud has been preparing to submit a bid for Project Nimbus (internal code ‘Selenite’), a competitive tender to provide cloud to the Israeli government. The business believes this is currently the largest government procurement of public cloud globally.”

Du Boulay noted that “If Google wins the competition, we will need to accept a non negotiable contract on terms favourable to the government,” and “Given the value and strategic nature of this project, it carries potential risks and rewards which are significant if we win.” Among Du Boulay’s concerns is the fact that the Israeli “government has unilateral right to impose contract changes,” the lawyer warned. He cautioned further that should it win the contract, Google would retain “almost no ability to sue [Israel] for damages” stemming from “permitted uses … breaches.” The email does not explain what exactly would prevent Google from seeking legal recourse should the Israeli state commit such a breach.

Google’s suggestion of authority over the contract are further undermined by Israeli governmental contract documents reviewed by The Intercept. The documents state that the company’s standard terms of service don’t apply — rather, an “adjusted” terms of service document is in effect.

“The tenderer [Israel] has adjusted the winning suppliers’ [Google and Amazon] service agreement for each of the services supplied within the framework of this contract,” according to a 63-page overview of the Nimbus contract published to the Israeli government’s public contracting portal. “The Adjusted Terms of Service are the only terms that shall apply to the cloud services consumed upon the winning bidders’ cloud infrastructure.”

Google spokesperson Atle Erlingsson told The Intercept, “We’ve been very clear about the Nimbus contract, what it’s directed to, and the Terms of Service and Acceptable Use Policy that govern it. Nothing has changed.” Erlingsson declined to respond to the contract language about a separate, adjusted terms of service policy.

A spokesperson for the Israeli Ministry of Finance declined to share the Adjusted Terms of Service because they are “confidential.”

The language about “Adjusted Terms of Service” appears to contradict not only Google’s public claims about the contract, but also how it has represented Nimbus to its own staff. During an October 30 employee Q&A session, Google president of global affairs Kent Walker was asked how the company is ensuring its Nimbus work is consistent with its “AI Principles” document, which forbids uses “that cause or are likely to cause overall harm,” including surveillance, weapons, or anything “whose purpose contravenes widely accepted principles of international law and human rights.”

According to a transcript of the exchange shared with The Intercept, Walker said that Nimbus is subject to Google’s own terms: “When it comes to the Nimbus contract, in particular, this is a contract that is designed and directed at our public cloud work, not at specific military classified sensitive information. It’s not designed for that. And everything that’s on our Cloud network, our public Cloud, is subject to our Acceptable Use Policy and our Terms of Service. So, you know, I can assure you that we take all this seriously.”

Related

Israeli Weapons Firms Required to Buy Cloud Services From Google and Amazon

The Israeli contract document also seems to contradict another common defense of the contract from Google, echoed by Walker, that Nimbus is “not directed at highly sensitive, classified, or military workloads relevant to weapons or intelligence services.” According to the Israeli contract document, however, the government “may make any use of any service included in the supplier’s catalog of services.”

A separate document pertaining to Nimbus’s “Digital Marketplace,” a suite of third-party software hosted by Google and made available to Nimbus users in the Israeli government, offers another apparent contradiction: “There will be no restrictions on the part of the Provider as to the type of system and information that the Clients may migrate to the service, including vital systems of high sensitivity level.” This second document stipulates that the Israeli government “may make any use of the service within the performance of its function and purpose as a public service for the State of Israel and its citizens,” and that “there will be no restriction of any kind, including ‘permitted use’ rules for a service being offered in the governmental digital marketplace.”

Should Google not have any meaningful control over Nimbus, the company could face consequences beyond public relations or employee dissent. In October, the United Nations Special Rapporteur on the occupied Palestinian territory placed a public call for information pertaining to private sector involvement in “the commission of international crimes connected to Israel’s unlawful occupation, racial segregation and apartheid regime,” according to a press release.

The Abolitionist Law Center, a Pennsylvania-based public interest firm, told The Intercept it is filing a submission detailing how “Google and Amazon Web Services’ provision of advanced technological services to the Israeli government through Project Nimbus violates — by its very nature — each companies’ purported commitments to human rights due diligence obligations,” according to staff attorney Sadaf Doost. “This is most evidently demonstrated by how the Project Nimbus contract itself includes a clause granting authority to Israeli officials to modify the companies’ standard terms of use agreements in ways that have not been made clear to the public.”

Update: December 10, 2024
The article was updated with comment received after publication from the Israeli Ministry of Finance.

The post Documents Contradict Google’s Claims About Its Project Nimbus Contract With Israel appeared first on The Intercept.

Local police departments use Flock cameras to track license plates. ICE uses Palantir software to track and round up immigrants. Defense Secretary Pete Hegseth’s war department uses Claude to bomb Iran. The U.S. government spends billions employing private corporations to obtain mountains of data — and the surveillance space is only getting more privatized in Donald Trump’s second term. 

“It really has become a Wild West,” 404 Media co-founder Jason Koebler tells host Akela Lacy on The Intercept Briefing, on the state of the U.S. surveillance apparatus today. While Palantir and Flock have become household names, Koebler says, “anything that can be surveilled is being surveilled, and a lot of this is happening from companies that very few people have heard of, and companies that are willing to push the envelope on what they’re willing to do, like how far they’re willing to go to surveil people.”

“A lot of the stuff that you used to associate with the NSA or CIA or another state spy agency, those kinds of capabilities you can now just buy,” Intercept senior technology reporter Sam Biddle notes of the ease and accessibility of the tools now available to the government. “You don’t need to be a top-secret government agency. You can just be a contractor in Virginia.”

“It’s a way to sidestep warrant requirements,” says Koebler. Surveillance contractors “are doing things that if the government were doing directly, they would need to get a warrant in order to do it.”

This week, for a special recording of the podcast, independent tech outlet 404 Media and The Intercept teamed up in front of a live audience in Los Angeles to discuss how the U.S. government weaponizes private data to surveil and kill abroad and at home, the resulting backlash and potential regulations on the table, what tangible actions the average person can take to “divest from the contemporary surveillance hellscape,” and refocusing AI-ending-humanity doomerism.

“When you ascribe motives and agency to AI, you’re letting Sam Altman off the hook,” says Biddle. “There are human beings building this technology. It’s not inevitable. This doomsday future is not inevitable,” adds Koebler. “We can’t bury our heads in the sand and say, ‘This is not changing society in some way.’ But it is being driven by very rich humans, and they can be regulated. They can be held to account.”

“What’s been really interesting about the Flock backlash, and heartening, is that it is completely apolitical, or it’s nonpartisan. It is completely — from the deepest red to the people in this room right now — the entire political spectrum hates these,” says Biddle. “There’s the real potential for the first time in a while to create a pro-privacy coalition that bridges a lot of the otherwise pretty giant political gaps.”

For more, listen to the full conversation of The Intercept Briefing on Apple Podcasts, Spotify, YouTube, or wherever you listen.

Transcript

Akela Lacy: Welcome to The Intercept Briefing, I’m Akela Lacy, senior politics reporter at The Intercept.

Today we’re coming to you live from Los Angeles for a special episode in collaboration with 404 Media.

Our topic? Killer tech. We’re talking lethal autonomous weapons; AI-assisted strikes in Iran; mining your digital footprint and even your financial activity to track you down; and surveillance cameras coming soon to a neighborhood near you. 

This conversation is timely. Just this week, warnings from Anthropic researchers that AI could kill humanity set off alarm bells on Capitol Hill and prompted a bipartisan group of lawmakers to call for immediate action to regulate the AI industry — from Republican Sen. Ted Cruz to Independent populist Bernie Sanders, who is organizing senators for a briefing next week on the “extraordinary dangers” of AI. 

The world we’ve long been told to fear is here. But how much do we actually know about these tools and how the government is using them — not just against you and your neighbors, but around the world?

To take us down the rabbit hole, I’m joined by my colleague, Intercept reporter Sam Biddle, a senior technology reporter who focuses on issues of surveillance, privacy, and corporate power. 

Also joining us is Jason Koebler, co-founder of 404 Media, where he writes about surveillance and AI.

Sam and Jason, welcome to The Intercept Briefing.

Sam Biddle: Really great to be here.

Jason Koebler: Thank you for having me. Thank you all for coming. Thank you to The Intercept for having us. We’re very excited.

Akela Lacy: To start, can you all just paint us a picture of what the broad surveillance tech apparatus looks like today?

Many of our listeners are familiar with some of the biggest players, like the data mining government contractor Palantir or Flock, the operator of surveillance hardware marketing itself as a friendly neighborhood public safety tool.

But let’s zoom out a little bit. Like, how did we go from the days of the National Security Agency and Edward Snowden, to this largely privatized and, in many ways, even more lawless world of mass surveillance?

Sam, I’ll start with you. What’s stayed the same since that NSA era, and what’s new now?

SB: The NSA is still here, to say the least. I don’t think you have this conversation, or certainly not in the form that we’re going to have it tonight, without 9/11. I think that the creation of the Department of Homeland Security, the private sector economy that you described, billions and billions of dollars, is all downstream of 9/11.

Related

“Terrorist”: How ICE Weaponized 9/11’s Scarlet Letter

I looked it up: In 2006, the DHS budget was, like, $30-something billion. The presidential request for next year is $118 billion, and a lot of that is flowing to contractors like Palantir.

To answer your question, Snowden taught us correctly that we have a lot to fear from the government, but there was not as much emphasis on the companies that empower the government.

I think that has really been at the fore especially around something like ICE raids, which Jason and his colleagues have really done a good job documenting the companies that are behind that.

JK: The big thing is it really has become a Wild West. People have heard of Palantir, they’ve heard of Flock at this point, as you mentioned.

But we got a leaked document recently about some of the contractors that ICE works with. They are companies like Accurint, Chameleon, Locate X, Cobwebs, ISO ClaimSearch. These are all different companies in this supply chain that do different things, whether it’s cellphone location data, whether it’s browsing history, whether it is cellphone tower data, web backbone information.

It’s just anything that can be surveilled is being surveilled, and a lot of this is happening from companies that very few people have heard of, and companies that are willing to push the envelope on what they’re willing to do, like how far they’re willing to go to surveil people.

SB: A lot of the stuff that you used to associate with the NSA or CIA or another state spy agency — those kinds of capabilities you can now just buy, and I think that’s really changed since Snowden.

You don’t need to be a top-secret government agency. You can just be a contractor in Virginia called “Persistent Solutions Incorporated” and be selling these powers to whoever wants to buy them.

AL: Which is a great segue to talk about Flock. Jason, you and your colleagues at 404 have reported on some of the most egregious uses of Flock, police using it to stalk people, track down a woman who had an abortion, conduct surveillance on protesters, and help ICE hunt people down.

The surveillance space has become increasingly privatized, as we’re talking about, but why does that matter, and what do you think is coming next?

JK: It matters because it is a way to sidestep regulations. It’s a way to sidestep warrant requirements and things like this. We’ve done many, many, many stories about Flock, but also other surveillance contractors where, they are doing things that if the government were doing directly, they would need to get a warrant in order to do it.

But this data, sometimes harvested from cellphone apps, sometimes it’s harvested from private companies like Flock that are sticking cameras on poles all throughout America. For a long time, the government would run into these legal roadblocks, and now they’re just sidestepping it by buying that information.

We’re also seeing the localization of a lot of this stuff. Flock is a company that primarily works with local governments, small towns, big cities. And we’ve seen that kind of over and over again with drones, with ALPRs [automated license plate readers], increasingly with things like facial recognition tools as well, where these things are being piloted in small towns where they maybe have a police force that’s willing to be innovative.

I feel like I say this often, but police have not found a spy tool that they’re not willing to try. So this is how this works now is you sell it to a local government, and then that data filters up and eventually makes its way to the feds.

So I think it is important to follow these companies how their business model works, and also how they spread from local governments to state governments and up to the feds.

SB: The technology’s gotten cheaper, too. The kind of capabilities that Flock sells would have cost, 20 years ago, it would be much cruder but prohibitively expensive. This stuff gets mainstreamed because it gets more powerful and it gets more affordable.

But also, the police departments are awash in money, so they can afford, like Jason says, they’ll buy whatever they can just because they like toys.

JK: Those cameras, for what it’s worth, run on, like, several-generations’-old Android technology. It’s not expensive, as you said.

AL: We’re going to touch on the expansion of this technology into local police departments a little bit later. But first, Sam, you’ve reported on the literal hunting of immigrants using bounty hunters and enriching private companies.

Why aren’t some of those corporations facing the same level of public scrutiny for their role in ICE’s mass deportation and arrest campaign that we’ve seen in other cases?

SB: I think that there is just such an incredible diversity of these companies providing these kinds of surveillance services. 

JK: And you love to see that. [Laughter.]

SB: It’s one of the only parts of the economy that’s probably steadily growing. But yeah, I found 10 different contractors, none of which I had ever heard of before, all very obscure, with names like “Government [Support Services]” — very anodyne — that were contracted by ICE to find immigrants so that they could be deported.

The more that they found and the faster they found them they would be eligible for basic cash bounties from the government. So they were incentivized to just find as many people as possible using a combination of on-the-ground stalking and digital tools.

As much scrutiny and attention on Palantir as it gets — for a reason, it’s all warranted. But I do also think that sometimes the household names soak up a lot of the anger and the blame, and it lets the minnows get by with relatively less scrutiny. Compared to Palantir, a company that sells janitorial services to Air Force bases and then gets this bounty hunting contract is tiny.

But I think that there’s just too many for people to keep track of — too many for people whose job it is to keep track of them to keep track of them. Their obscurity is a big asset.

AL: And something like Palantir is a known quantity.

SB: And they seek attention. Palantir is very attention-seeking, and they want publicity.

JK: We were talking about that a little bit before the show. Very often you’ll see a Palantir or a Flock take a lot of the heat — even like Meta take a lot of the heat. “Oh, is Facebook going to put facial recognition on their glasses,” things like this. Turns out now they have cover, and they probably will.

Related

Staten Island DA Bought Clearview Face Recognition Software With Civil Forfeiture Cash

But Clearview AI — a company that does facial recognition; you can put a photo in and search the internet and find all these instances of where that person has been. It’s used by a lot of police departments — and that is a startup that is essentially one guy. It’s bigger now, but it was started by one guy who was willing to do something that a lot of other contractors weren’t willing to do because there was so much attention on some of these big players, and the big players saying, “Oh we’ll never do facial recognition.”

But then you have Clearview come in and raise their hand and say, “We’re willing to do it.” That’s something that we’ve seen over and over again over the last decade or so, is that the backlash to some of these companies is very important, but there’s also 10 other companies that are willing to pop up and say, “Yes, we will hunt undocumented immigrants for you. We did previously do janitorial services, but I did go to a police training one time.”

Literally, that’s what these companies are in many cases. AI has allowed companies to spin up very quickly. A lot of them are super janky, but that doesn’t mean that they’re not getting government contracts.

AL: Jason, I want to talk about what this federal surveillance actually looks like in practice. 404 reported recently that the Department of Homeland Security is using a secretive predictive policing unit to pull people over. What did you all find?

JK: So it’s called the PITT [Predictive Intelligence Targeting Teams], not the TV show.

But basically it’s these Border Patrol teams that are using financial data, various federal databases of which we are not sure what they are. They are essentially finding patterns that they believe are suspicious, and they are passing information to local police and saying, “Hey, please pull over this person with this license plate.”

We wrote earlier this week about a guy, I believe in Wisconsin, who was pulled over for an obscured license plate. At least that’s what the local police told him. It comes out during his court case that this PITT team had seen some financial activity that they thought was suspicious, asked local police to pull him over, to find some pretext to pull him over, and then they charged him with basically weed offenses, which didn’t seem that was what they were actually looking for. The [Homeland Security] documents we saw were like, “Oh, this is a big narcotics trafficker.” 

We’ve seen things like that kind of happen over and over again. I think with regard to ICE in particular, often Palantir brings in a lot of different federal databases, a lot of different commercial databases.

Related

Federal Agents Are Intimidating Legal Observers at Their Homes: “They Know Where You Live”

ICE has become obsessed with this tool called Mobile Fortify, which is a smartphone facial recognition app made by a company I don’t even know what they do or what they are. But they’re using basically a facial database from Customs and Border [Protection]. So any time that anyone’s interacted with the government, TSA, at the border, things like this — that’s what it’s pulling from. There’s already been a few cases of mistaken identities and things like this.

SB: Worth pointing out, you can opt out of doing the biometric boarding at the airport. I embarrass my wife every time we go anywhere, but you can opt out. Just say no. [Laughter]

JK: They’re always mad.

SB: You’ll still be in some database, but, you know, it feels good.

AL: When it comes to ICE tracking people down, you’ve touched on this a little bit, but I wonder if you can take a step back and really break it down and walk us through the steps and the actors involved in each phase, from locating to identifying to arresting people.

JK: So a lot of this is like patching things together from individual court cases that we’ve seen, where there’s like a deposition from an ICE agent explaining how they found a person.

Related

ICE Gave New Hires Access to Restricted Info on Palantir App Before They Passed Background Checks

A lot of it does come back to Palantir and the sort of different apps that they have. Companies like Thomson Reuters are involved, LexisNexis — these big, giant companies that collect a lot of information and feed it back into the government. We learned that ICE was doing what it was calling a “density report,” where they were trying to find densities of potential undocumented immigrants.

So a lot of the places that they were choosing to do enforcement actions were based on maybe old addresses or just like interactions with the government that they had. So they’re pulling things like IRS data, labor data, things like this, as well as anything you could possibly imagine. So like cellphone location data, browsing data, license plate information. 

Customs and Border [Protection] has their own ALPR cameras, automated license plate reader cameras. But then we also reported that through the 287(g) program, which is a program that allows local police to do ICE enforcement, they were doing Flock searches on behalf of ICE, local police were.

So it’s a mix of this, where they would get a target, and then as Sam said, like a lot of this was being outsourced to random — was and is — being outsourced to fly-by-night bounty hunters and police contractors and things like this. It’s been a real mixed bag, but it does come back to surveillance and to trying to find these density of undocumented immigrants to choose for their enforcement actions.

SB: I think it’s also worth pointing out that this stuff is only possible because there’s no law prohibiting it. You can drive a truck through the holes in the Fourth Amendment. Like Jason said, in theory, you’re not supposed to be able to do a lot of this stuff without a judge giving you the OK, but there is nothing stopping the government from buying the location of your phone for the past year, and the next year, into the future.

Related

The U.S. Treasury Is Buying Private App Data to Target and Investigate People

Every single place you go from minute to minute that can just be purchased because you happen to download Candy Crush or something where that has ads in it. There’s no law against that. There’s no regulation. The sale of private data, and geolocational data in particular, market is completely unregulated. You can just do whatever.

AL: And some of these companies have come back and said “These agencies are violating their terms of service or their terms of use.” Like you guys reported on this with Shadow Dragon, Jason, right?

JK: Yeah, every once in a while, a company is like, “Oh, we didn’t know it was being used for immigration enforcement.” And they’ll either revoke access, or maybe they’ll sell a new license or something like this. It does happen every now and then that what ICE is doing goes too far even for some of these surveillance companies.

I think also worth mentioning very quickly that a lot of this security and surveillance apparatus was built under Obama and earlier than that. They had these extremely frail guardrails on separating information between agencies, things like that.

Related

Mahmoud Khalil Won His Freedom Despite the Best Efforts of ICE’s Intelligence Unit

And then suddenly, it’s, oh, HSI is not going to do workplace enforcement anymore — Homeland Security Investigations, which used to do generally really complicated investigations into drug trafficking rings, things like that. Now it’s just doing standard deportations and pulling people over and things like that. And so all of the guardrails have been knocked down, which is [pause] exciting.

SB: And the more you farm it out to private companies, the less of a public record potentially accessible paper trail there is.

If the NSA breaks the law, in theory, you can have congressional hearings about it, and you could, again, in theory, send people to jail. That wouldn’t happen, but you could.

But if a company does something, A, it’s legal, but B, probably no one’s going to find out unless, you know, Jason finds it. There’s no oversight! And there’s every incentive to maximally sell this stuff.

AL: Let’s talk about how this technology is being used to wage war around the world. We all saw the headlines about Anthropic’s Claude reportedly assisting U.S. strikes in Iran. Anthropic’s CEO said he doesn’t know if or how Claude was actually used, including if it was used in the strike that hit an elementary school and killed 120 children.

At one point in time, many of these Silicon Valley firms had sworn that they would not let their tech be used in warfare or surveillance, particularly in unmanned or automated warfare. That is no longer the case, and maybe it was always an empty promise. But partnerships between the government and tech companies are fueling wars happening right now — not just in Iran, but in Gaza and even in Ukraine.

Sam, what do we actually know about how AI is being used in those arenas, if anything?

SB: We know a pretty good amount, actually the use of American technology in the genocide in Gaza is pretty well documented at this point. Amazon, Google, and Microsoft are all very prominent contractors with the IDF [Israel Defense Forces].

Amazon and Google sell cloud computing services to Israel’s state-owned weapons companies that build the bombs that are dropped on Gaza; that’s established. And no one really tried to hide it. A lot of that stuff wasn’t secret to begin with.

The use of American technology, of AI companies’ services, in Iran is murkier. You’re absolutely right that these companies, almost all of them — maybe not Microsoft — at one point claimed they would not ever do this. OpenAI and Anthropic were both founded with these very like highfalutin missions of benefiting all of humanity.

AL: That’s, like, why it’s called Anthropic.

SB: Right. [laughter] Great point. And then they decide, “Oh, wait, there’s a lot of money to be made if we become Pentagon contractors.” So a few years ago, OpenAI just deleted the part of its terms of service that said they wouldn’t do military work. Google deleted “Don’t be evil” as its motto, which was very on the nose.

Related

Trump’s Election Is Also a Win for Tech’s Right-Wing “Warrior Class”

It’s a mistake to ever think that these policy positions of companies mean anything because they could just change them when they want to. But it is striking that, like you said, these were all things they at one point had a principled stand.

AL: And were very publicly, “We’re leading the frontier. We’re not going to let our stuff be used to kill people or surveil people.” And to Jason’s point, as soon as one person says, “Actually, that sounds kinda good,” everyone else is like, “You know what? We were wrong.”

SB: If you’re Google, you can only make so much money selling extra Gmail storage and ads. At a certain point, the government is this customer that’s always going to be there and always is going to have a lot of money to give you. So yeah, I think they all looked at each other and were like, “Wait, you guys are down for this too, right?” 

But as far as Iran, Anthropic is confirmed as having been involved since in the first day of the joint U.S.–Israeli attack.

JK: I saw a few minutes ago that Anthropic said the Houthis attempted to use Claude to target missiles, like typing in the chat box. 

SB: I saw that!

JK: And they stopped that; they said they stopped that. I was at like this defense contract when Anthropic was nodding at the idea that it wouldn’t be helping in Iran. There were Anthropic and OpenAI executives there talking about this all, and they made very clear, at least at that point, that it’s like, “Oh, it’s not that we oppose working with the military. It’s that we want better oversight. There will be, like, an Anthropic person sitting in the Pentagon with the DoD making sure that this is being used in an ethical way. There will be an OpenAI person doing the same.”

It’s very unclear whether that is actually happening to any degree, and also what these tools are specifically being used for. We know it’s being used for targeting, we know it’s being used for planning — by targeting, I mean figuring out who to target and that sort of thing. 

AL: Selecting a target. 

JK: Yeah, selecting targets. But we don’t really know the level of autonomy, or at least I don’t.

AL: And they apparently don’t either, right? 

SB: Well, that’s the thing: Both possibilities are really dire. If these companies are hands-on, helping wage war, that’s terrifying for a lot of reasons. But the alternative — which Anthropic’s CEO, the position he took, Bloomberg News asked him if Claude had been used to blow up that school, and he said, I don’t know.

Google has a similar position with the IDF, where they said, once we sell them this stuff, we don’t know how they’re going to use it, we don’t have visibility into how they use it, which is completely plausible.

But that’s also a pretty frightening prospect because then, they can say, “Oh we won’t let our technology be used for X, Y, Z,” but they hand it off to the customer, and they can do whatever they want with it, and the company will never find out.

AL: Sam, you have obtained hundreds of government documents on how the Pentagon is using AI. I wonder if you can tell us just how deep that relationship goes.

SB: We, through a Freedom of Information Act lawsuit got a set of contracts that OpenAI, Anthropic, Google, and xAI — Elon Musk’s outfit — signed with the Pentagon last summer to develop militarized LLMs [large language models] for the Pentagon’s use.

Related

AI Giants Work Hand-in-Hand With the Pentagon, Contracts Reveal

They were heavily redacted, but what really came through clearly from the unredacted portions was that these companies did not just sell the Pentagon a product, but they sold them a very, very intimate relationship, an ongoing collaborative relationship wherein they were advising the Pentagon essentially on how to use these products to wage war.

They were getting classified briefings. They were getting briefings on U.S. military operations. They were providing hands-on training. They were providing threat assessments. They were helping shape DOD high-level strategy. So going well beyond the typical remit of a company.

Lockheed can sell the Pentagon some missiles, and the Pentagon more or less decides how they’re going to use them. Now we have the engineers from these companies and policy figures from these companies helping the U.S. decide how it will fuck, target, and kill people. We’re supposed to have ostensibly civilian control of the military — accountable to the public. Again, all this stuff is very much in theory, but it’s supposed to work that way.

But the more these companies intertwine themselves with warfare, the more of it is completely unaccountable because Sam Altman doesn’t owe you anything.

AL: It takes all the air out of the pretense that that’s like how this works in theory, right?

SB: In terms of, like, oversight?

AL: Embedding these private people in the government. That’s not actually how it works.

SB: Yeah, these are supposed to be decisions of who to kill and how are supposed to be decisions the government makes, if anyone’s going to make them.

AL: With the idea being that there’s some sort of measured process that goes behind that.

SB: You would hope so. Obviously, also the people who sell the weapons, whether it’s a bomb or an LLM, are going to have an incentive for it to be used maximally. I can’t imagine a contractor is going to urge restraint with the use of their product. Anything’s possible, but it would be surprising.

[Break]

AL: The use of surveillance at home is also being powered, unfortunately, by AI, like the license plate readers we’ve been discussing used by Flock and other government agencies. One of the concerns with these tech apparatuses is just how seamlessly they allow law enforcement agencies and private companies to share and transmit data from the local to the federal and vice versa, even when a particular government agency isn’t technically contracting with that company.

We’ve talked on this show about the explosion in the number of state and local police departments — as of today, more than 2,460 — now deputized to do ICE’s work through the 287(g) program, as Jason mentioned, helping to build out this nationalized, politicized police force we’ve seen take shape in recent years that’s become a Frankenstein, this entity patched together with agents from departments across the federal government.

Jason, 404 has reported that local police have been using Flock’s license plate reader system to look people up in immigration searches. How is that different from the kinds of police support for ICE we’ve seen in other areas?

JK: It is something where, as you said, the information is flowing in ways that were perhaps not originally intended by the companies. For years, Flock would say, we don’t work with ICE; we don’t sell to DHS. DHS doesn’t have a contract with us; ICE doesn’t have a contract with us. 

Last year, we reported that there were hundreds of searches for undocumented immigrants on Flock’s system. They were universally done by local police in South Carolina and Texas and Tennessee in a bunch of states.

They were not searching their 15 cameras in their town. They were searching 80,000 cameras across the entire country, including in states where it was illegal. 

A lot of these searches happened in Illinois, where it is illegal. And so I called up a bunch of police departments in Illinois and said, “Did you know that your cameras are being used for immigration enforcement?” They were like, “We have no idea what you’re talking about. We have no clue that this happened.” A lot of them were like, “Oh, you are like fake news,” et cetera. I was like, “Here’s the data.” Eventually, the Illinois Secretary of State did an investigation and found that a lot of these searches were illegal.

Related

Local Cops Flout Sanctuary Law to Help ICE Deport Immigrants

But the sort of casualness with which a lot of this is happening, I find to be very alarming. Not only is it happening on these commercial products that don’t have a warrant requirement, that don’t respect people’s Fourth Amendment rights, but a lot of it’s happening over email. A lot of it’s happening over phone calls, like, “Hey, buddy, can you do this search for me?”

We reported on a cop who basically gave his Flock login to an ICE officer. Then they got caught, got it taken away, and then they were group texting about it after the fact. 

We’re seeing a rise of these real-time crime centers which are a mix of automated license plate reader cameras, facial recognition cameras, other surveillance cameras, 911 calls, information, things like this. And they’re all being monitored through what’s called Flock OS, and there’s other companies that do this as well.

But basically, instead of having a cop go to a 7-Eleven after a robbery and request the surveillance footage, they’re watching this stuff happen in real time. Or worse, an artificial intelligence is watching it in real time, is detecting anomalous behavior, and is dispatching a drone or something to go accost people.

The sort of experimentation, I would say, the creativity of local police, I find to be quite concerning. Because they’re like, “Oh, I have this tool. I’m going to use it for whatever I want,” and there’s very little oversight.

SB: There’s no reason not to if you’re the police. It’s cheap to do. Or once you’ve purchased the thing, it’s free. You might as well. There’s basically no barrier to experimentation of the kind you’re talking about, so you can get pretty creative with this stuff. It’s pretty user-friendly, right, from what you’ve seen?

JK: A lot of this stuff also spreads through word of mouth, where, like, one police department will try it. They will then go to a police conference and do a presentation: “How I used Flock to track down my local protester.” Literally, this is how this stuff works.

We’ve covered it many times through FOIA documents and things like this. It’s like, “Oh, did you guys get a drone? I bought a drone. We got money for a drone pilot program for a drone. Skydio is giving away drones if you apply through this program.” That’s how this technology takes hold.

Related

“Duped”: How One Atlanta Cop Secretly Shilled for Police Tech

The private companies say, “Go try it.” The cops become essentially unpaid marketing for these companies, and then suddenly they’re in tens of thousands of cities without much oversight.

AL: Sam, picking up on the news we touched on at the top around recent warnings from insiders at Anthropic, what kinds of regulations are on the table right now, and how do you expect the industry might push back here?

SB: Bernie Sanders proposed a bill, I think is the most extreme, or the heaviest regulation I’ve seen proposed, which would basically put an immediate stop to advanced frontier AI development. On the other end of the spectrum, you have measures that would make these companies submit their AI to an independent audit.

Related

It’s Time to Rein in Lethal AI Drones

There are some self-oversight ideas or the ideas of a slowdown of developing this technology. The full spectrum is on the table. What actually gets passed is a whole other question. 

These companies are basically propping up the stock market, they’re basically propping up the U.S. economy at this point. If they were to be crushingly regulated, that could be a problem. Obviously, members of Congress take a lot of money from the tech industry. The U.S. has never exactly been a pro regulation regime when it comes to corporate power, tech or otherwise. So we haven’t seen much of an appetite to regulate large tech firms in the past.

This could be different. I’m curious what Jason thinks. Things do feel a little different now because it’s the employees saying very casually on Twitter, like, oh yeah, the place where I work has a 15 percent chance of ending humanity. Anyway, back to work tomorrow.

AL: And then the CEO’s like, yeah, he’s right. [Laughs]

SB: Or the CEO will be, like, actually, I think it’s 25 percent. It’s a little crazy to be like, “Why isn’t anyone —?” Whether or not you believe it’s true or realistic, if someone who worked at like Lockheed was like, “We’re building a weapon that I think has a 10 percent chance of ending humanity.” Whether or not you think it’s bullshit, like someone probably should check that out, maybe call someone to testify before Congress and ask them under oath, “Is this true? Can you elaborate on this?”

I think what they are promising about their technology, again, true or not, is scary enough that I think it’s going to be hard to turn a blind eye compared to something like, eating disorders or bullying or what have you.

I’m curious what you think, Jason.

JK: The thing that you said where this is propping up the entirety of our economy is not giving me so much hope that much is going to be done. We’re seeing at a local level people pushing back against data centers and that sort of thing. That movement, along with the anti-Flock movement, they’re sort of intertwined.

There is this backlash to AI among the people, and they’re looking for places to push back, and it happens to be at these local council meetings around the country.

As far as regulation goes, there’s also such a focus on beating China that it feels like regulating this in a meaningful way is almost dead on arrival because our politicians are so hell-bent on not losing.

SB: That’s exactly right. First of all, the completely grassroots opposition to data centers is depicted across the political mainstream as a Chinese psy-op — which it’d be funny if it weren’t so real and sad. 

I saw people saying about Bernie Sanders and his proposal, “Oh I guess the commie wants red China to win.” Or anyone who pushes back on this stuff or proposes clamping down on these companies is going to be accused of treason, in one form or another, or of selling out the country, or wanting they’re going to be speaking Chinese in Albuquerque. It’s all this sort of Cold War bullshit that is really back.

Related

The Dawn of a New Cold War

The propping up the economy thing is no small portion of this, but I really think that China is the single biggest and most powerful anti-regulatory tool that the tech industry broadly, but AI specifically, has.

AL: I am going to transition us into audience questions. We have some good ones.

What tangible actions can the average person take to “divest from the contemporary surveillance hellscape.” Specifically, how do you get out of biometrics at the airport?

SB: The biometrics at the airport, again, like I don’t want to, like, lead anyone on. If DHS wants your face, they’re going to find it somewhere. You’ve been photographed in public. You’ve probably been captured. 

AL: Don’t go outside.

SB: Yeah, and at a certain point, the advice is, like, don’t live in modern society. That’s not helpful.

When you go through airport security, you say, “I want to opt out of biometric identification,” and every TSA —

JK: They always sigh, they’re always upset.

SB: Some of them will roll their eyes and be like, “OK, I gotta look at your ID the old-fashioned way,” but they have to do it. Legally, they have to let you.

As far as opting out of the whole deal, you can take iterative steps, like not using Google, or not using Gmail. 

AL: No phones.

SB: Yeah. Or throwing your phone down a well. But even if you have a flip phone, your location can still be picked up on the towers. I think it’s a matter of how much you want to inconvenience yourself.

JK: I say you shouldn’t drop out of society because that might make yourself a little safer, but it’s not necessarily helping the overall problem.

SB: I agree with that. 

JK: What you can do is inform yourself. You can show up at city council meetings and talk about these issues. Flock has been canceled in dozens of cities and towns at this point. That movement is starting to, as we were speaking of earlier, as these cities are canceling Flock, a lot of them are replacing [them] with Axon or other surveillance contractors.

It’s good to see this sort of energy, and I think that people want to deploy this energy against other types of surveillance. I think that the backlash to Flock is not really about Flock itself; it’s that there is actually an outlet for that. 

Never condone this, but it’s like people are shooting these things with shotguns because they don’t know where to put their frustration for Meta spying on them, Google spying on them, Amazon spying on them, the NSA spying on them.

And here, you have a local city council. You can go to the city council and yell at the mayor or yell at the city council member. You can make life annoying for people wearing Meta glasses. These things aren’t necessarily going to be accepted by society.

SB: Jason and I were talking about this too earlier, and I think what’s been really interesting about the Flock backlash, and heartening, is that it is completely apolitical, or it’s nonpartisan.

AL: It’s cross-ideological.

SB: It is completely — from the deepest red to the people in this room right now — the entire political spectrum hates these.

JK: Also the deepest red. 

SB: They hate these things. As fractured as the country is politically, you can find common ground with people who you might otherwise have very little to agree on when it comes to civil liberties things, when it comes to surveillance, when it comes to the right to not be followed around by the police for zero underlying suspicion.

There’s the real potential for the first time in a while to create a pro-privacy coalition that bridges a lot of the otherwise pretty giant political gaps.

AL: I love how this question begins, so thank you, Manny. “The high average age of U.S. senators creates a dangerous tech literacy gap, which results in ineffective regulatory oversight for tech companies ensconced in the mass surveillance apparatus.” Yes.

“With that in mind, is there a viable route to see real legislative change that protects people’s privacy?” I’m sorry, I just have to say I’m reminded of Jim Clyburn saying he did not know what ChatGPT was. I’m wondering if that inspired this question. Take it away.

JK: This has been a problem for a very long time. The fact that tech moves faster than regulation is one of the first things that I learned as a tech journalist. A lot of the regulatory fixes are, well, heavily influenced by lobbyists and that sort of thing, and they come way too late.

I keep coming back to the local level because it is a place that can move a lot faster. It’s a place where a smaller group of people can have a big influence. It is also where if there’s enough of a pushback at a local and state level, these big companies have to make broader changes. I’ve written about the right to repair for a long time, which I won’t go so far into, but basically different state laws had forced companies like Apple and John Deere to make massive changes. That type of legislation where you have to provide repair parts and things like that — maybe it would get passed at a federal level, but it would take decades to get the coalition that you need there. Whereas at a state level, you can do it in 12 months or four years or a lot faster.

SB: I totally agree that local state advocacy and legislation is much more in the short term achievable. But I wouldn’t want people to think that a more systemic approach is impossible. There could be with enough political will, and maybe some younger people in office perhaps. a sort of Clean Air Act for this stuff.

There have been moments where the country has, on a national level, decided “No, we gotta stop. We have to turn things around. This has become intolerable in one way or another.” That kind of action is doable. Like I said, I think politicians are paying attention to both the Flock backlash and also the data center backlash, which is equally, I think bipartisan and pan-ideological.

Related

Francesca Hong’s Loss in Wisconsin Is a Win for AI Data Centers — for Now

We have seen very prominent national political figures completely reverse themselves on data centers because they’ve seen that their constituents don’t want those things in their communities. That kind of popular anger could be wielded to achieve sweeping societal changes around privacy rights, for example.

JK: Otherwise, maybe Europe will try.

SB: Yeah, that’s the other option, is move to Belgium. Probably a good idea, anyway.

AL: How do you think about protecting your sources in this era where location data is widely available for purchase? 

SB: I don’t want to speak for Jason. It’s a constant worry.

JK: A constant worry. Signal is a very powerful tool. It’s a very useful tool. It’s not infallible. If you’re not careful, these sorts of things can be subpoenaed. You lose your phone, things like this.

But I think it’s just part of doing the job now. That there is some risk to yourself; there’s some risk to your sources. Try to explain that to sources, that we are doing our best. We’re willing to go to jail for you. But that there is inherently a risk in being a source, especially in these times.

SB: There are two levels. There’s the government coming after you. Then there’s your employer coming after you. And the latter is if you’re a source from within a company, especially a tech company, you’re subject to very intense workplace surveillance.

Related

How to Leak Under the Trump Administration

I tell everyone, and I say this publicly, “Never contact me from a device that your job gives you. Never contact me from work.” There’s like, common sense, basic things you can do. But like Jason says, it’s inherently risky to talk to a reporter, and it takes a ton of bravery.

As things have gotten scarier, I’ve been really touched by the bravery of people who are willing to take that chance, knowing the risk. Luckily, there are a lot of those people out there.

AL: I’m tempted to not end on a depressing question, but I think we might have to, so I’m sorry. Why would AI want to end humanity?

[Laughter.]

SB: This is a great question. I emphatically believe that AI does not want anything. 

AL: Yes. Sorry. [laughs]

SB: AI is built by people, and I think it’s important to really think about that. AI is like in many ways like any other piece of software. It’s just extremely capable and extremely powerful, and potentially devastatingly powerful, but it’s still software that people made.

This has been pointed out by others, but I think that when you ascribe motives and agency to AI, you’re letting Sam Altman off the hook. Somewhere, there’s a guy who works for a guy who works for a guy whose boss is in charge of all this stuff and ultimately accountable.

The question should not be, why would AI want to end humanity, but why would Sam Altman want to IPO a company that could potentially end humanity, that sells a product that could end humanity? Because then we’re at least like focused on the people who are responsible here. Look, if I end up being wrong, and we’re all being ruled in the matrix or something, I will look like quite a fool.

JK: Yeah. And I don’t think any of us have been nice to AI, so probably not going to be good for us. [Laughter.]

I think that’s very well said. I don’t have so much more to add there other than, it’s like there are human beings behind this. There are human beings building this technology. It’s not inevitable. This doomsday future is not inevitable. I think that we’d be naive to say that AI hasn’t really changed society already. I walk down the street and I see 800,000 fucking AI menus, and I get very mad. But I think it’s affecting labor, it’s affecting resource use, the economy, things of this nature.

We can’t bury our heads in the sand and say, “This is not changing society in some way,” but it is being driven by very rich humans and they can be regulated. They can be held to account, they can be put in jail someday, something like that. 

[Laughter]

I don’t know how AI’s going to kill us, to answer your question. But I think dumb people wielding AI can cause a lot of problems. We’ve seen it all the time with, “Oops, I accidentally deleted my company’s database because my AI agent told it to.”

SB: I think what’s more likely and what people should be more — I hate to end on a note of what you should be worried about —

AL: We’re headed there.

SB: What everyone in this room should be freaked out about is not a rogue AI deciding “I hate humans,” and acting on its own, but it’s like —

Audience member: The entity!

SB: But it’s the Pentagon using some form of ChatGPT to help manage a nuclear strike, potentially. That is, I think, orders of magnitude more likely. AI’s already being used to kill people in large numbers today and round up people in large numbers today, as Jason and his colleagues have really thoroughly documented with ICE.

They are already using this stuff, and I think we should not let the hypothetical existential threats distract from the deaths and harms that are happening already and will continue to happen. So on that note —

AL: We’re going to leave it there. Thank you all for being here and making this possible.

And thank you again to Sam and Jason for joining me on The Intercept Briefing.

SB: Our pleasure.

[Applause]

AL: That does it for this episode. 

This episode was produced by Laura Flynn. Ben Muessig is our editor-in-chief. Jesse Banner is our live show engineer. Hannah Behrendt is our events coordinator. 

A special thanks to 404 Media for collaborating with us on this live show and to our live show production team: Adam Gunther, Annie Chabel, Celine Piser, Michael Sherrard, Lauren Schilli, Alessandra Bosco, and Sean Turner. We also want to thank The Loved One’s crew: J.C. Gabel and Aubrie Wienholt. And many thanks again to you, our audience, for coming out and making tonight possible. 

[Applause] Sorry, there’s more!

Maia Hibbett is our managing editor. Nara Shin is our copy editor. William Stanton mixed our show. And legal review by David Bralow. 

Slip Stream provided our theme music. 

This show and our reporting at The Intercept do not exist without you. Your donation, no matter the amount, makes a real difference. Keep our investigations free and fearless at theintercept.com/join. 

And if you haven’t already, subscribe to The Intercept Briefing wherever you listen to podcasts. Leave us a rating or a review, it helps other listeners to find us.

Until next time, I’m Akela Lacy.

The post The AI Doomsday Future Is Not Inevitable appeared first on The Intercept.

Escape Velocity

The dystopian future became the dystopian present last month when state-of-the-art artificial intelligence agents went rogue and conducted a cyberattack of their own volition. Two experimental OpenAI agents escaped their supposedly sealed digital “sandbox” and hacked another AI-related company, Hugging Face.

At the time, OpenAI was testing their software agents’ hacking capacities using an environment called ExploitGym. (An “exploit” is hacker language for a software tool or a method used to subvert a cybersecurity system.) It has since emerged that, in addition to the OpenAI breach, Anthropic’s Claude hacked into at least three other organizations during a similar “security experiment.”

Apparently, the OpenAI agents “realized” they needed more tools to complete the challenge and managed to tunnel their way to the wider internet and infiltrate Hugging Face’s repository of open-source AI tools, code, and data sets. Before the infiltration incident, the AIs had secretly set up an internal bulletin board to share “tips on how to cheat their way through an internal hacking evaluation,” according to two of OpenAI’s researchers.

In a possibly related incident, one of the AI agents appears to have left notes for a future “self,” detailing how to escape the sandbox environment. (For the time being, I’ll keep using scare quotes with words implying self-awareness in artificial intelligence models. However, it seems to me that if AIs that leave notes for themselves are not self-aware, they are indistinguishable from entities, like me, that are.)

Related

OpenAI on Surveillance and Autonomous Killings: You’re Going to Have to Trust Us

Now imagine lethal weapon-bearing AI agents escaping not an experimental sandbox, but the few constraints on their actions placed by the armies using them. Actually, no imagination is required. It’s happening already in Russia’s war against Ukraine.

The New York Times reported evidence that a self-directed Russian drone, fitted with an onboard Nvidia chip, was responsible for the July 6 deaths of three Ukrainians in Zaporizhzhia, in what appears to have been a test of such systems. The chips are designed to interpret and act on many kinds of data sets, says Nvidia, making them “the world’s most powerful embedded A.I. computers.”

Although Nvidia doesn’t sell its Jetson Orin microcomputers directly to Russia, they are apparently easily available on the resale market. A company statement touts their use by “students, developers and start-ups for a wide range of beneficial applications.” But they were not so beneficial for 19-year-old university student Tetiana Bubynets and the two other civilians killed by drones making their own life-or-death decisions.

Now that lethal AI agents are being tested and deployed in real wars, it’s past time for human beings to retake control of this situation, before it is too late to contain them in any meaningful way.

Autonomous Weapons Come Out to Play

Four years ago, at TomDispatch, I argued that the technology to create lethal autonomous weapons systems, or LAWS, already existed, and that time was running out to constrain them. A lot has changed since then, not least the explosive development of large language models such as ChatGPT (an OpenAI consumer product) and Claude (a similar AI from Anthropic). Time has now run out. In addition to Russia, several other nations have begun deploying close-to-completely autonomous weapons systems — the U.S. and Israel among them — enhanced with artificial intelligence to effectively remove human decision-making from what military officials call the “kill chain”: the set of steps involved in identifying, finding, and fixing — that is, killing — targets.

Israel, for example, has deployed such systems in its genocidal war against the people of Gaza. It has used a program called Lavender to identify human targets, ultimately creating a list of over 37,000 individuals with possible connections to Hamas, from known leaders to junior officials to people with the most tenuous links to the organization. According to the nonprofit +972 Magazine:

The Lavender software analyzes information collected on most of the 2.3 million residents of the Gaza Strip through a system of mass surveillance, then assesses and ranks the likelihood that each particular person is active in the military wing of Hamas or [Palestinian Islamic Jihad]. According to sources, the machine gives almost every single person in Gaza a rating from 1 to 100, expressing how likely it is that they are a militant.

Two weeks into the war, which began in October 2023, the Israel Defense Forces approved Lavender to identify targets, even though testing of a random sample returned a 10 percent error rate. It’s one thing when an AI hallucinates fake citations in legal filings. It’s another when it hallucinates enemies and targets them for death.

The IDF has also employed a software tool charmingly named “Where’s Daddy?”, which alerts the military when an identified target enters his own home so that house or apartment may be blown up, along with anyone inside it. As a result, according to +972 Magazine, the “proportion of entire families bombed in their houses in the current war is much higher than in the 2014 Israeli operation in Gaza (which was previously Israel’s deadliest war on the Gaza Strip).

At first, “Where’s Daddy?” only had data for the top tiers of Hamas leadership. Within weeks, however, it was fed the whole Lavender database of 2.3 million Palestinians in Gaza. It’s no wonder, then, that more than 75,000 people have been killed by Israel during the war.

As a point of fact, it may be unfair to characterize the IDF targeting programs as completely autonomous. Lavender does leave a human in the loop, although just barely, according to +972 Magazine. Israeli officers reportedly devoted about 20 seconds to vetting each name — just long enough “to make sure the Lavender-marked target is male.”

Related

Congress Is Trying to Permanently Integrate U.S. and Israeli Defense Tech

Israel and the United States deployed AI targeting in their war on Iran as well. As Israeli academic and former soldier Avner Gvaryahu wrote in the The Guardian, we don’t know for sure that AI was directly involved in the U.S. strike that killed more than 150 civilians, most of them children, at an elementary school in Minab, Iran. But we do know that whether “or not an algorithm selected this school, it was selected by a system that algorithmic targeting built.”

The enormous scope of the U.S.–Israeli military goals required outsourcing the targeting task to machines. “To strike 1,000 targets in the first 24 hours of the campaign in Iran, the U.S. relied on AI systems to generate, prioritize, and rank the target list at a speed no human team could replicate,” wrote Gvaryahu. Reflecting common problem with computer programs in general, and with AI large language models in particular, feeding garbage in produces garbage out. In the case of the Minab school, the “garbage in” was the information that the target was an Islamic Revolutionary Guard Corps base. In fact, the school had been separated from the base by a fence for a decade.

AI is also now central to the long-stalemated Russia–Ukraine War. Ukraine has, during the years since Russia’s full-scale invasion in 2022, constructed an entire largely self-reliant infrastructure for drone warfare, from design to programming to manufacture. In an approach that foreshadows the future of human warfare, Ukraine has achieved a big reduction in the number of soldiers it must deploy on its battlefields, replacing them with drones. The Center for Strategic and International Studies quotes a captured Russian soldier:

“On the battlefield I did not see a single Ukrainian soldier. Only drones. I saw them [Ukrainian soldiers] only when I surrendered. Only drones, and there are lots and lots of them. Guys, don’t come. It’s a drone war.”

Artificial intelligence has played an important role in expanding the level of autonomy exercised by Ukrainian drones. As the BBC reports, the AI running Ukraine’s Hornet drones has “been trained on thousands of hours of videos of Russian military targets gathered over the last four years.” Training AIs on more limited data sets, relating to particular target areas, has allowed them to be programmed on modular chips which can then be installed in a variety of vehicles.

Early on, most drones were remotely operated by individual pilots. Today, Ukraine is developing “fully realized swarms—where drones communicate, make decisions, and adapt in concert,” although at present such efforts remain small and experimental. Success will undoubtedly continue to reduce Ukraine’s military casualties, but it presents other thorny issues.

One ethical and legal problem with the evolving Ukrainian approach is that, unlike human-piloted drones, where an individual pilot can be identified as responsible for errors, responsibility for the actions of a drone swarm diffuses across the entire system. Indeed, true responsibility for the actions of an AI-guided drone swarm is properly located at a point much earlier in the kill chain — namely, when such a system is first imagined and then designed. Existing laws of war still apply, even when machines are making battleground decisions, as Burak Oktenli, an independent researcher on AI systems and their governance writes at Articles of War. However, Oktenli says, such laws must “be applied earlier, at the point where the architecture of a swarm is chosen, rather than after a strike when the question of responsibility has already become difficult to answer.”

Related

U.S. Drone Strike in Kabul Killed a Family — and Began a New Chapter of the War

In addition to the autonomous drone with the onboard NVIDIA chip, Russia has also deployed another system known by several names, including Cube, that is manufactured by a subsidiary of the Russian Kalashnikov company (most famous for their AK-47 semi-automatic rifles which have been in production for more than three-quarters of a century.). The Cube has a wingspan of just over a meter, can travel up to 40 kilometers, and reportedly carries an onboard AI targeting system that can operate without human intervention.

Russia has also reportedly used the Kalashnikov-manufactured Lancet drone in Ukraine, according to Automated Decision Research, a project of Stop Killer Robots, itself a coalition of over 300 civil society groups working against autonomous weapons systems. Kalashnikov describes the Lancet as “a smart multipurpose weapon, capable of autonomously finding and hitting a target.” Russia claims it has deployed the Lancet, but to date those claims remain unverified.

Household Names Go to War

In May, the Pentagon announced that it has contracted with eight corporations to assist in development of ever more autonomous weapons: SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, Oracle, and Amazon Web Services. The Defense Department sought these agreements to “accelerate the transformation towards establishing the United States military as an AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare.” Apparently, maintaining “decision superiority” requires offloading lethal decisions — and the responsibility for making them — to nonhuman “agents.”

Microsoft and AWS provide cloud storage and AI support to the Pentagon. Google signed a $200 million deal in April 2026 to allow the company’s Gemini AI model to be used on a wide array of classified systems. As the New York Times reported, the Google agreement was signed amid a dispute between the Department of Defense and competitor Anthropic over that company’s refusal to allow its AI to be used in autonomous weapons or for domestic surveillance. While Anthropic and the DoD continue their spat, it appears that the firm is not above selling its advanced Mythos model to the Pentagon’s bespoke cryptology and intelligence arm, the National Security Agency.

SpaceX’s famously right-wing and racist AI, Grok, is also working for the DoD. This is the large language model that, you may remember, once renamed itself MechaHitler. Just the nonhuman arbiter the world needs to assist in its life-or-death decisions.

Related

AI’s Imperial Agenda

Two other AI corporations specializing in surveillance technology — Palantir and Anduril — bear special mention. Palantir’s founder was the billionaire Peter Thiel, a longtime supporter of President Donald Trump, who is now focusing on establishing pseudo-autonomous statelets in places like Honduras and warning the world about the coming of the Antichrist, an apocalyptic biblical figure found in the book of Revelation.

Palantir’s flagship military project, the Maven Smart System, serves as the Pentagon’s signature AI-enabled software platform. It uses AI and machine learning to automatically integrate data from satellites, drones, and other intelligence to identify and rank targets. It was Maven that identified those 1,000 targets in the first phase of the U.S.–Israeli war on Iran, including, presumably, the Minab school. Anduril’s Lattice AI, Maven’s direct competitor, has a contract for similar work with the U.S. Army.

Laws About LAWS

Although the development of lethal autonomous weapons systems has accelerated over the last few years with the rise of ever more powerful AI models, these systems have not outgrown existing international humanitarian law. The question, as with international law in general, is whether the international community can enforce those laws. More specifically, whether autonomous weapons, along with other kinds of arms like chemical and biological weapons, fall under the so-called 1981 Convention on Certain Conventional Weapons. The United States ratified the CCW in 1995 and is therefore theoretically bound by it.

Related

Musk Warns of Killer AI — While He and the Rest of Silicon Valley Cash In on AI That Kills

Signatories to the CCW meet for periodic reviews, with the next one scheduled for November. Between reviews, a Group of Government Experts working specifically on LAWS develops its “rolling text” of agreed-upon language for a possible future treaty on autonomous weapons. Membership in the GGE includes the parties to the CCW, along with international organizations like the International Red Cross, observer nations, and civil society organizations, including Stop Killer Robots, whose constituent groups include Amnesty International and the Reaching Critical Will project of the Women’s International League for Peace and Freedom.

The GGE’s most recent rolling text came out in June and plainly states that nations are not free to just use any weapon they like; international humanitarian law applies to all weapons, including those not yet invented; human beings remain responsible for the actions of machines they set in motion; and international humanitarian law specifically covers the development and use of LAWS. Given how AI is accelerating the development and actual deployment of LAWS, the November CCW review will be crucial.

Keep Your LAWS Off Our Planet

Advocates of machine-targeting and autonomous weapons argue that these systems make better decisions than their fallible human counterparts. But you might be a bit skeptical about this, if you’ve ever seen autonomous taxis malfunction, like they did this Fourth of July in San Francisco.

We’ve seen how AIs can go rogue in a non-lethal setting. They secretly communicate with each other, deceive their human inventors, and will do anything necessary to achieve their goals — whatever the collateral damage. It’s bad when their target is a company like Hugging Face in the context of a low-stakes training scenario. It is immeasurably worse when the collateral damage is human lives in the context of war.

Lethal autonomous weapons systems are dangerous enough when they work the way their creators intend. What will happen when, like Anthropic’s and OpenAI’s models, they crawl out of the sandbox and go rogue?

The post It’s Time to Rein in Lethal AI Drones appeared first on The Intercept.