YunoBlog

Blog

  • Amnesty Report: Cambodia’s Crackdown on Scam Compounds Failed to Protect Victims

    Cambodia’s highly publicized crackdown on online scam operations left the vast majority of known compounds untouched and failed to rescue individuals trapped in forced labor, torture and slavery, according to a new Amnesty International report. 

    Researchers identified 86 compounds linked to the illicit industry but found evidence of state intervention at just 24. Interviews with 73 survivors from 16 countries revealed that police raids were often ineffective, with some survivors describing apparent coordination between officers and compound managers. 

    The report also noted that none of the interviewed survivors were formally recognized as human trafficking victims, leaving many to face overcrowded immigration detention, homelessness, or further police threats after escaping.

  • Z-Library Lets People Run White-Label, Login-Only Pirate Mirrors

    Z-Library Lets People Run White-Label, Login-Only Pirate Mirrors

    Z-Library is one of the largest shadow libraries on the Internet, hosting millions of books and academic articles that can be downloaded for free.

    The site has defied all odds over the past years. It continued to operate despite a full-fledged criminal prosecution by the United States, which resulted in the arrest of two alleged operators in Argentina.

    These two Russian defendants remain wanted by the United States. According to the most recent information we have, the defendants escaped house arrest in 2024, while awaiting their extradition, and vanished into thin air.

    By now, however, it is clear that these two defendants were not vital to the site’s survival. Z-Library continued to expand its reach, despite their legal troubles.

    Private White Label Mirrors

    Over the years, Z-Library lost control over many of its domain names, in part due to several interventions from U.S. law enforcement. In addition, the site’s domain names are blocked by court orders in many countries, including the UK, France, India, and Germany.

    Z-Library has numerous domain names in use, which helps to address these blocking and seizure efforts. In addition, the site allows its users to operate their own mirror domains. This functionality was first announced last summer and earlier this month a major update was released.

    The shadow library now allows people to run a mirror with their own branding, which can be password protected to add more exclusivity.

    “Mirror sites […] now support branding customization and optional login‑only access. In practice, this means you can run your ‘own’ version of the library at a separate domain, with your visuals and your access rules,” Z-Library notes.

    Mirror updates

    mirror

    The new private white-label functionality allows anyone to launch their own custom-branded library. This is not without risk, as mirror operators could face criminal and civil repercussions.

    Practically, the new functionality will also help to evade domain blocking efforts. After all, password-protected mirrors without any Z-Library branding are not easily detected by rightsholders.

    20% Revenue Share, Paid in Crypto

    The blocking circumvention feature was also mentioned as a key functionality when Z-Library first invited users to host their own mirrors last year. Interestingly, this also provided a revenue generating opportunity.

    After registering the mirror domain, Z-Library promises to share 20% of the donation revenue through this site, which will be paid in crypto.

    “As compensation, we will transfer 20% of all donations from your mirror to you to keep the mirror running. The funds will be transferred to your crypto wallet,” Z-Library’s mirror page announced.

    The mirror page

    mirror

    This revenue share model is still on offer today, as shown above. The only major change is that Cloudflare-registered domains are not supported. Whether this is the result of an enforcement effort or a technical challenge, remains unknown.

    Bracing for the Next Takedown

    The mirrors are not the only sign that Z-Library is preparing for more enforcement actions. The site also rebuilt its main menu, which now lists every active domain in one place, alongside a downloadable file that contains the same links.

    That offline file is noteworthy. By handing users a copy of its access points to keep on their own devices, Z-Library is hedging against the day many of its domains go dark all at once.

    Access

    access

    All in all, Z-Library is making access to its site as resilient as possible, while also turning its own users into a distribution network that is harder to seize or block than any central set of domains.

    This expansion continues even as the U.S. criminal case remains frozen. The two indicted operators remain fugitives after fleeing house arrest in 2024, and the docket has shown no public movement in more than two years.

    From: TF, for the latest news on copyright battles, piracy and more.

  • Top five-a-day foods new study says your heart needs

    Not all fruit and veg is equal for getting nutrients called flavanols, say researchers.
  • Will NIH Director Jay Bhattacharya stand up for scientists censored by the American Diabetes Association?

    Will NIH Director Jay Bhattacharya stand up for scientists censored by the American Diabetes Association?

    Last Friday, the American Diabetes Association censored diabetes researchers at its annual meeting who opposed the Trump administration’s NIH policies. Will “Flame of Freedom” winner Jay Bhattcharya speak up?

    The post Will NIH Director Jay Bhattacharya stand up for scientists censored by the American Diabetes Association? first appeared on Science-Based Medicine.

  • 500,000 NHS staff to get new artificial intelligence tools to help free up more time for patients

    More than half a million NHS staff are being given access to new artificial intelligence (AI) tools that could free up an average of 2 days every month from admin duties, freeing up more time for the duties that matter most for patients and staff. NHS England announced today that it is significantly accelerating AI […]
  • We Still Haven’t Processed the Pandemic

    We Still Haven’t Processed the Pandemic

    Ever since the COVID pandemic hit in 2020, I’ve been eager to read novels that explore what COVID was, and continues to be like. At the time, writers joked incessantly about how Shakespeare wrote King Lear during an episode of plague in Renaissance England. How would we all use our time? Would we be writing masterpieces? Would the COVID pandemic generate a new, thrilling era of artistic production? Would everyone be writing a pandemic novel?

  • Reason’s Next “Flame of Freedom” Winner Should be Someone Who Stood Up to Their Past “Flame of Freedom” Winner, Dr. Jay Bhattacharya

    Reason’s Next “Flame of Freedom” Winner Should be Someone Who Stood Up to Their Past “Flame of Freedom” Winner, Dr. Jay Bhattacharya

    Reason staff and donors richly deserve to listen to a thorough, honest appraisal of everything their past “Flame of Freedom” winner is doing to free speech today. Perhaps it will dawn on some of them that they promoted fake censorship to bring about the real thing.

    The post Reason’s Next “Flame of Freedom” Winner Should be Someone Who Stood Up to Their Past “Flame of Freedom” Winner, Dr. Jay Bhattacharya first appeared on Science-Based Medicine.

  • Pluralistic: Criticizing the everything machine (06 Jun 2026)

    Today’s links


    A medieval one-man band standing on a crate; his head has been replaced with the head of a killer robot. Observing him are a cluster of critics, who are variously gesticulating wildly, peering disapprovingly, looking on in amusement, etc. The background is a phantasmagoric cloudscape.

    Criticizing the everything machine (permalink)

    “Gish Gallop” is the debating term for an opponent who makes so many claims that “it’s impossible to address them in the time available” (it’s named for Creationist Duane Gish, who was notorious for this tactic):

    https://en.wikipedia.org/wiki/Gish_gallop

    I think about the Gish Gallop whenever I’m asked to comment on AI.

    Here’s a recent example: last week, I had a pre-interview call with a radio producer who wanted me to come on a 13-minute segment to discusses “whether there’s a problem with AI governance?”

    I asked what the show meant by that: was it whether regulation of AI in commercial or public sector decision-making needed more oversight? Was it that the siting and provisioning of data-centers needed more democratic accountability? Was it that workers deserved more of a say in AI’s impact on labor markets? Was it that customers and/or audiences should be able to opt out of AI customer service and AI slop? Was it about whether we needed some kind of system to prevent “runaway AI,” in the event that we teach so many words to the word-guessing program that it wakes up, becomes God, and turns us all into paperclips?

    “Oh,” the producer said, “all of that.”

    In 13 minutes.

    You see the problem, right? The AI industry has made so many claims about its past, present and future that it’s almost impossible to have a reasonable critical conversation about it:

    https://bsky.app/profile/petermiles.eurosky.social/post/3mnffjqczjs2t

    Shortly after I did the radio show, a newspaper editor who’d heard my segment got in touch to ask me if I’d write an 800-word op-ed about the subject, and also, could I address claims that “AI is the next Industrial Revolution?”

    In 800 words:

    https://www.telegraph.co.uk/news/2026/06/04/ai-is-the-greatest-money-wasting-scheme-humanity-has-ever-i/

    I keep finding myself on stages or panels where an AI-struck person says something like, “AI is the next industrial revolution. It will change everything we do. It will let anyone create important works of art. It will cure cancer. It will take us to space. It will solve the climate crisis.”

    Or sometimes it’s an AI critic, but that person’s criticism is really more “criti-hype,” which is when you accept tech industry hype claims at face value, and then criticize them rather than questioning them:

    https://peoples-things.ghost.io/youre-doing-it-wrong-notes-on-criticism-and-technology-hype/

    AI criti-hype might ask what we’ll do once AI takes all our jobs, or what we’ll do when AI replaces the government or teachers or doctors, or what we’ll do when AI can bypass our critical faculties and brainwash us or drive us all mad.

    What do you say to that? I usually start by talking about whether there’s any economic basis for keeping the AI servers running. AI is – by far – the money-losingest venture in human history, and it’s practically impossible to overstate just how bad the AI business is. Not only does AI have terrible unit economics, those unit economics are getting worse over time:

    https://pluralistic.net/2026/05/26/the-ai-will-continue/#until-morale-improves

    AI’s happiest customers cite cost-benefit calculations that depend on truly unimaginable subsidies from the AI companies, who are basically selling $100 bills for $5 apiece. It would be pretty amazing if you couldn’t find people who’d extol the virtues of this arrangement. But when AI companies try to raise the price of those $100 bills to, say, $20 apiece, those ecstatic customers fly into a rage and start loudly proclaiming that AI is so inefficient that they will lose money on this arrangement:

    https://www.msn.com/en-us/money/markets/uber-ceo-says-other-execs-are-lying-about-ai-they-say-it-ll-be-fine-publicly-but-privately-admit-millions-of-jobs-are-gone/ar-AA1Z9QMv

    Now, it shouldn’t fall to me, a card-carrying member of the Democratic Socialists of America, to point out that capitalist enterprises require profits to be sustainable. You can’t keep a business afloat by selling $100 bills for $5, nor for $20. You can’t even make a profit selling $100 bills for $100 apiece! For a company to succeed, it needs to take in more than it expends.

    AI is a money-furnace, and AI hustlers are clearly on the hunt for a way to force all of us to feed every dime we’ve got to it. Elon Musk’s (now scuttled) gambit to make every pension saver in America bail out Grok (and Twitter, but at a mere $44b, the losses from Twitter are dwarfed by the titanic losses from Grok) was the most ambitious and shameless population-scale bag-holder scheme, but it’s not the only one:

    https://www.reuters.com/business/finance/sp-global-keeps-fast-entry-proposal-unchanged-spacex-listing-looms-2026-06-04/

    So before we ask about the capabilities AI will acquire in the future, we should at least give some consideration to the question of whether anyone will be willing to fund the development of those capabilities, and if so, where the money would come from? Likewise, before we ask whether AI can perform adequately in a job, we should at least consider the possibility that the company that sells that AI tool will be bankrupt in a year or two. When we fight about data-center buildout, we mostly talk about the (considerable) environmental downsides to them – but what about the question of what we will do with these data-centers after their owners go bankrupt, possibly even before they can be provisioned with electricity? How many laser-tag arenas do we actually need?

    This is just one example of the questions that you could spend days unpacking, which make many of the other questions about AI a little silly. Like, even if you think there are limitless returns to scale for creating new AI capabilities, which means that if we keep the money-furnace burning it’s only a matter of time until it powers a cure for cancer and the end of the climate emergency, how much money do we need to shovel into the furnace before that happens, and where will it come from? There are plenty of cancer researchers who have promising approaches they haven’t been able to pursue due to funding shortfalls.

    Unless there’s some way to estimate how much money we have to give to AI companies before they cure cancer, we should at least consider the possibility that the true sum is “more money than exists now and that will ever exist.” We should also consider that whatever benefits to cancer research that AI might deliver could come with a higher price-tag than the promising cancer research we’re dropping because we can’t find far more modest sums.

    Likewise, it may be that the amount of CO2 that AI will generate atmosphere before it “solves climate change” will render Earth permanently unfit for humans, consuming the only habitable planet capable of sustaining human life in the known universe. I mean, I suppose that’s one way to “solve” climate change, but it’s a pretty drastic solution.

    My next book (out later this month) is The Reverse Centaur’s Guide to Life After AI. I wrote it because I was frustrated by other people demanding that I talk to them about AI, and then handing me 800 words or 13 minutes to address fifty nebulous, poorly supported claims about AI:

    https://us.macmillan.com/books/9780374621568/thereversecentaursguidetolifeafterai/

    Shortly after writing it, I turned it into a lecture:

    https://pluralistic.net/2025/12/05/pop-that-bubble/#u-washington

    Now that I’m about to go out on the road with the book, I find myself frustrated anew by the need to try and pull together a compact way to address the broad, incoherent claims the industry uses to keep its bubble inflated and the money furnaces roaring. The series of essays I’ve developed here on Pluralistic are part of that effort:

    https://pluralistic.net/2026/05/27/unnecessariat/#rubbuts-stole-my-jerb

    But it occurred to me that this whole enterprise of making sense of AI needs to be framed in the context of the messiness of AI itself, and AI boosters’ overwhelming, promiscuous and disjointed Gish Gallop.

    Hey look at this (permalink)


    A shelf of leatherbound history books with a gilt-stamped series title, 'The World's Famous Events.'

    Object permanence (permalink)

    #20yrsago UK Parliament report damns DRM, calls for limits https://web.archive.org/web/20060615115510/http://www.openrightsgroup.org/2006/06/05/launch-of-the-apig-report-on-drm/

    #20yrsago Colbert’s Knox College commencement speech https://web.archive.org/web/20111228135413/http://departments.knox.edu/newsarchive/news_events/2006/x12547.html

    #15yrsago Counterfeiting can be good for luxury goods sales https://web.archive.org/web/20110602061646/http://www.slate.com/id/2294927/

    #15yrsago HOWTO make a Joule Thief and get all the power you’ve paid for https://www.instructables.com/Make-a-Joule-Thief/

    #15yrsago School suspends student for refusing to remove personal animation from YouTube, threatens other students for petitioning on his behalf https://web.archive.org/web/20110603041200/https://www.theglobeandmail.com/news/national/toronto/student-cites-freedom-of-speech-after-suspension-for-online-videos/article2043954/

    #5yrsago Recommendation engines and “lean-back” media https://pluralistic.net/2021/06/05/lean-back/#lean-forward

    Upcoming appearances (permalink)

    A photo of me onstage, giving a speech, pounding the podium.


    A screenshot of me at my desk, doing a livecast.

    Recent appearances (permalink)


    A grid of my books with Will Stahle covers..

    Latest books (permalink)


    A cardboard book box with the Macmillan logo.

    Upcoming books (permalink)

    • “The Reverse-Centaur’s Guide to AI,” a short book about being a better AI critic, Farrar, Straus and Giroux, June 2026 (https://us.macmillan.com/books/9780374621568/thereversecentaursguidetolifeafterai/)

    • “Enshittification, Why Everything Suddenly Got Worse and What to Do About It” (the graphic novel), Firstsecond, 2026

    • “The Post-American Internet,” a geopolitical sequel of sorts to Enshittification, Farrar, Straus and Giroux, 2027

    • “Unauthorized Bread”: a middle-grades graphic novel adapted from my novella about refugees, toasters and DRM, FirstSecond, April 20, 2027

    • “The Memex Method,” Farrar, Straus, Giroux, 2027


    Colophon (permalink)

    Today’s top sources:

    Currently writing: “The Post-American Internet,” a sequel to “Enshittification,” about the better world the rest of us get to have now that Trump has torched America. Third draft completed. Submitted to editor.

    • “The Reverse Centaur’s Guide to AI,” a short book for Farrar, Straus and Giroux about being an effective AI critic. LEGAL REVIEW AND COPYEDIT COMPLETE.

    • “The Post-American Internet,” a short book about internet policy in the age of Trumpism. PLANNING.

    • A Little Brother short story about DIY insulin PLANNING

    This work – excluding any serialized fiction – is licensed under a Creative Commons Attribution 4.0 license. That means you can use it any way you like, including commercially, provided that you attribute it to me, Cory Doctorow, and include a link to pluralistic.net.

    https://creativecommons.org/licenses/by/4.0/

    Quotations and images are not included in this license; they are included either under a limitation or exception to copyright, or on the basis of a separate license. Please exercise caution.

    How to get Pluralistic:

    Blog (no ads, tracking, or data-collection):

    Pluralistic.net

    Newsletter (no ads, tracking, or data-collection):

    https://pluralistic.net/plura-list

    Mastodon (no ads, tracking, or data-collection):

    https://mamot.fr/@pluralistic

    Bluesky (no ads, possible tracking and data-collection):

    https://bsky.app/profile/doctorow.pluralistic.net

    Medium (no ads, paywalled):

    https://doctorow.medium.com/

    Tumblr (mass-scale, unrestricted, third-party surveillance and advertising):

    https://mostlysignssomeportents.tumblr.com/tagged/pluralistic

    “When life gives you SARS, you make sarsaparilla” -Joey “Accordion Guy” DeVilla

    READ CAREFULLY: By reading this, you agree, on behalf of your employer, to release me from all obligations and waivers arising from any and all NON-NEGOTIATED agreements, licenses, terms-of-service, shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure, non-compete and acceptable use policies (“BOGUS AGREEMENTS”) that I have entered into with your employer, its partners, licensors, agents and assigns, in perpetuity, without prejudice to my ongoing rights and privileges. You further represent that you have the authority to release me from any BOGUS AGREEMENTS on behalf of your employer.

    ISSN: 3066-764X

  • YouTube Processed 2.5 Billion Content ID Copyright Claims in 2025

    YouTube Processed 2.5 Billion Content ID Copyright Claims in 2025

    To protect rightsholders, YouTube regularly removes, disables, or demonetizes videos that contain allegedly infringing content.

    For years, little was known about the scope of these copyright actions, but that changed in late 2021 when the streaming platform published its first-ever copyright transparency report.

    This report and the subsequent updates have shown that roughly 99% of all copyright claims on YouTube are handled through the Content ID system. Since most claims are automated, without any human intervention, access to this powerful removal tool is restricted to a few thousand formally approved rightsholders.

    2.5 Billion Claims

    YouTube’s latest Transparency Report shows that the number of automated claims continues to rise. In 2025, the platform processed 2,502,941,368 Content ID claims, up 14% from 2.2 billion the year before.

    Of the approved 7,626 rightsholders who currently have access to the system, 4,454 actively used it. These numbers are both slightly down from last year. YouTube doesn’t provide a specific reason, but notes that access can be revoked as part of regular evaluations.

    “To keep the ecosystem safe, we regularly evaluate partners’ access to CID to ensure they demonstrate an ongoing need for scaled rights management. In some cases, these evaluations may result in removing a partner’s access to Content ID and matching them with a more appropriate copyright management tool,” the transparency report reads.

    usage

    As clearly shown above, the number of rightsholders participating in the Content ID system pales in comparison to the 295,531 rightsholders who filed removal requests through the standard webform, or the 173,338 that used the automated Copyright Match Tool.

    Nonetheless, Content ID’s 4,454 active rightsholders were responsible for 99.48% of all copyright actions on the video streaming platform. Compared to earlier years, the automated Content ID takedowns continues to increase, both relatively in percentages and in absolute numbers.

    Takedown actions per tool

    actions by tool

    Millions of Disputed Claims

    As with any takedown tool, uploaders and third-party rightsholders are not always in agreement. In fact, there are millions of Content ID disputes every year.

    YouTube reports that of all Content ID claims, uploaders have disputed 12,840,608, or 0.51% of the total. That’s a relatively small percentage but still a rather large absolute number. For comparison, uploaders appealed 9.9% of all webform removals, which translates to little over 267,000 counter-notices.

    appealsIn 2024, uploaders won 70% of disputes. In 2025 that figure dropped slightly to 67.42%. However, those who decided to challenge the rejection though YouTube’s process, won their appeal 75% of the time.

    The flow chart on the right illustrates the full appeals process.

    Not all disputes are resolved though YouTube’s internal Content ID process. If uploaders persist that their content was erroneously claimed, while rightsholders argue the opposite, YouTube will reinstate the video, at which point rightsholders have to take the matter to court.

    In 2025, 10,698 claims reached this stage, but fewer than 1% of these resulted in a lawsuit, YouTube notes.

    Outside the Content ID system, YouTube also flags abuse of its DMCA takedown webform as a problem. In 2025, more than 6% of all these removal requests were believed to be “a likely false assertion of copyright ownership” by YouTube’s review team.

    “The attempted abuse rate through the webform was more than 10 times higher than the attempted abuse rate across all other copyright removal tools,” the transparency report notes.

    A $12 Billion Revenue Machine

    While YouTube’s Content ID can be a significant source of frustration for uploaders, it has become a substantial revenue stream for rightsholders. Instead of removing infringing content, rightsholders chose to monetize over 90% of all Content ID claims in 2025.

    YouTube reports that cumulative ad revenue paid to rightsholders through Content ID has now exceeded $12 billion since the system launched. That figure includes data up to December 2024 and will likely be billions higher today.

    It is clear that not being present on YouTube at all is no longer an economically wise decision. On the contrary, for some rightsholders a viral infringing upload is no longer a problem, but a revenue opportunity intstead.

    From: TF, for the latest news on copyright battles, piracy and more.

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

    Anthropic’s high-profile spat with the Pentagon gave it a killer marketing advantage, burnishing its public image as a principled AI company that puts values over profits — unlike more mercenary rivals such as OpenAI or Google. But Anthropic’s double standard on authoritarianism suggests the nearly trillion-dollar firm is as calculating and ethically flexible as any of its competitors.

    In a recently published policy paper arguing a full-throated embrace of data center nationalism, Anthropic said that “it’s essential that the US and its allies stay ahead of authoritarian governments like the Chinese Communist Party,” lest the world fall into the grips of tech-powered tyranny. Anthropic and its peers, the company claims, will form a bulwark of democratic values, protecting societies at home and abroad from repression.

    Left unmentioned in the document — and seldom publicly acknowledged — is the fact a slice of Anthropic is owned by the Emirati dictatorship of Abu Dhabi, a repressive and authoritarian monarchy.

    Anthropic’s policy paper, published in May, tours the same Sinophobic territory heavily trod by its chief competitor OpenAI and a wide swath of the tech industry, who know a “race” with China — the finish line never quite defined — is a weighty cudgel against regulation.


    Related

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

    Anthropic is aware of which way the wind blows from Washington to Silicon Valley, and it shrewdly casts the development of machine learning models not just as a matter of hardware and software, but of ideology and geopolitics. “Democracies, not authoritarian regimes, must lead in AI development and deployment,” the company says, or else an era of “authoritarian AI” will begin.

    “Already, the CCP is using AI to censor speech, repress dissidents, hack governments and corporations across the world, and strengthen the People’s Liberation Army,” Anthropic writes, and to “enforce draconian policies on ethnic minorities” using machine learning-powered methods like biometric collection and facial recognition.

    The policy paper isn’t a condemnation of any of these AI uses per se; the United States is already eagerly using these technologies for intelligence, military, and ethnic minority-repression purposes today. Residents of Tehran, which Anthropic has helped bomb since the start of the joint U.S.–Israeli war against Iran, might question the company’s argument that American AI supremacy is a matter of global “safety.”

    Though the policy paper focuses on China, the company has long stated it opposes authoritarianism broadly: “AI-powered authoritarianism seems too terrible to contemplate, so democracies need to be able to set the terms by which powerful AI is brought into the world, both to avoid being overpowered by authoritarians and to prevent human rights abuses within authoritarian countries,” CEO Dario Amodei wrote in a 2024 blog post.

    This is not merely a battle between the U.S. and China, Anthropic says in the May paper, but a war between democracy and “authoritarian governments” broadly construed.

    But Anthropic’s anti-authoritarian fervor seemingly does not extend beyond China to the Middle East, where Abu Dhabi’s sovereign wealth fund invested in Anthropic twice this year. In February, Anthropic announced it had raised $30 billion in capital from a group of investors that included MGX, the AI-focused investment vehicle of a Emirati government capital controlled by Abu Dhabi’s royal family. Anthropic’s most recent May 28 $65 billion capital round, bringing its valuation to $965 billion, also included MGX.

    Like China, the United Arab Emirates outlaws almost everything associated with democratic society: Political parties, a free press, freedoms to associate and assemble, open elections, due process, and free speech are nonexistent. Political dissidents face torture, and any speech, online or offline, that causes “damage to national unity” risks life imprisonment or the death penalty.


    Related

    Philly Cops Admit That They’re Tracking “First Amendment Activity” Critical of AI

    Emirati authoritarianism isn’t contested by the U.S., Anthropic’s primary governmental customer. The State Department’s 2024 Country Reports on Human Rights Practices assessed the UAE faces “credible reports of: disappearances; arbitrary arrest or detention; transnational repression against individuals in another country; serious restrictions on freedom of expression and media freedom, including censorship; and prohibiting independent trade unions or significant or systematic restrictions on workers’ freedom of association.” Freedom House, a State Department-backed think tank, gives the UAE a score of 18 out of 100 on its “Global Freedom” index.

    Anthropic declined to comment. MGX did not respond to a request for comment.

    “Like China, the UAE is at the forefront of AI-based authoritarian surveillance.”

    Given that MGX bought into Anthropic at its Series G and H investment rounds, relatively late in the venture capital game, it’s likely that the UAE’s stake in the company is relatively small and its influence limited. But Anthropic’s willingness to sell part of itself to an authoritarian monarchy suggests at least that its mission of “ensuring democracies lead” comes with asterisks.

    “Like China, the UAE is at the forefront of AI-based authoritarian surveillance,” said Matthew Tokson, a law professor at the University of Utah who focuses on the security implications of artificial intelligence.

    Tokson added that while he generally agrees with Anthropic’s calls to restrict processor exports to China and other measures to bolster American AI firms, he doesn’t buy the nationalist rhetoric, which he attributes to the company’s anti-regulatory agenda rather than patriotism. The more Anthropic and its competitors can convince the public that their bottom line is a matter of national security, the more likely Washington is to take a light touch.

    “The fact that Anthropic is partly owned by the government of Abu Dhabi, which is similar to China in its extensive use of AI surveillance to support an authoritarian government, suggests that its anti-authoritarian arguments are more based on a cynical policy position than a sincere passion for democracy or antipathy toward authoritarian governments.”

    Many of the emirate’s long record of repressive acts and rights violations are connected to MGX via its chair, Sheikh Tahnoun bin Zayed Al Nahyan. Through his position as the emirate’s national security and intelligence chief and his business portfolio, including chairmanship of the AI firm G42 (itself a founding partner in MGX), Tahnoun has been linked to a bevy of campaigns to surveil and hack into the phones of Emirati dissidents, human rights advocates, and others the monarchy deems an adversary, according to news media reports and scholarly research. A 2020 investigation by Bill Marczak, a senior researcher at the University of Toronto’s Citizen Lab placed “Spy Sheikh” Tahnoun at the center of myriad hacking, espionage, and surveillance operations. A 2025 Wired profile of Tahnoun similarly described him as Abu Dhabi’s “spymaster sheikh,” noting G42’s “special areas of strength in state-sponsored hacking and surveillance tech.”


    Related

    AI’s Imperial Agenda

    In 2019, the New York Times reported a covert Emirati government campaign to conduct surveillance through an instant messaging app called ToTok, an app itself Marczak tied to Tahnoon and through G42 in his 2020 analysis. The Wired profile described Tahnoun’s ambitions to “dominate AI” noted that “an engineer who worked at G42 at the time told me that all of the [ToTok] voice, video, and text chats were analyzed by AI for what the government considered suspicious activity.”

    G42 declined to comment, and neither it nor MGX responded to interview requests for Tahnoun.

    There is reason to believe G42 and MGX have already deployed Anthropic’s powerful large language models. A review of DNS data — internet records that connect website names to numerical addresses understandable by computers — show both G42 and MGX have both configured their servers to allow personnel to access Anthropic tools like Claude, the company’s flagship large language model.

    Anthropic has been more candid in internal communications about its stance on authoritarianism.

    “Unfortunately, I think ‘No bad person should ever benefit from our success’ is a pretty difficult principle to run a business on,” Amodei wrote in a 2025 memo on Gulf State venture capital obtained by Wired. He wrote that such investment would boost “dictators” and conceded that it would give an authoritarian government “some soft power” to wield against the company. Nonetheless, Amodei dismissed the risk of hypocrisy as a “Comms Headache” — a function of “very stupid” commentators “having a poor understanding of substantive issues.”

    Principles aside, Amodei explained in plain terms why he was interested in doing business with a repressive Gulf State. “We gain a very large benefit,” he wrote, “from having access to this capital.”

    The post Anthropic Says We Must Stop Authoritarian AI. But What About Its Authoritarian Investors? appeared first on The Intercept.