UN Clash: Altman and Amodei Warn of AI Catastrophe—As Trump Calls It a Hoax
Altman and Amodei Tell the UN What Trump Wouldn't Hear: "No Level of Catastrophic Risk Is Acceptable"
TL;DR
- On September 23, 2026 - one day after Trump told the UN General Assembly that AI safety concerns were a "globalist scheme" and a hoax - Sam Altman (OpenAI) and Dario Amodei (Anthropic) briefed the UN Security Council directly, arguing almost the opposite.
- Altman: "If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone." Amodei: "If managed poorly, I even believe that AI could be a risk to humanity as a whole."
- Yoshua Bengio, co-chair of the UN's AI panel, rejected the "race" framing outright: "The race is not a law of nature. It is the product of choices, choices made by the companies themselves."
- This isn't purely a war of words. Fortune reported that Trump has separately named Amodei directly in a Truth Social post threatening "tremendous CRIMINAL and REGULATORY power," and that the Pentagon had already designated Anthropic a supply-chain risk, stalling release of its Mythos model - meaning the disagreement already has real consequences attached to it, not just competing speeches.
Reviewed by Imran Khan Pathan, Editor at AI Tech Safar. Every direct quote here is checked against at least two named outlets - AP, Bloomberg, Axios, and Fortune's coverage all overlap closely on the wording, which gives me real confidence in them. The Pentagon/Anthropic detail specifically comes from Fortune's reporting and isn't repeated elsewhere in the pieces I checked, so I've flagged it as a single-sourced claim rather than treating it as independently confirmed twice over.
Last updated: September 23, 2026.
48 Hours, Two Completely Different Rooms
On September 22, Trump stood before the UN General Assembly and called AI safety warnings a hoax, rejected any "globalist scheme" to control the technology, and said the US would keep its lead over China "by a lot" - the full story is in our Super Intelligence piece. Less than 24 hours later, the two most powerful people in commercial AI sat in front of the UN Security Council and made close to the opposite case.
The session was organized by France, which holds the Security Council's rotating presidency this month, specifically to address the danger of AI misuse and the possibility that advanced models could slip beyond human control. Altman attended in person; Amodei joined remotely by video, alongside Hugging Face co-founder Clément Delangue and Yoshua Bengio, the Turing Award-winning co-chair of the UN's Independent International Scientific Panel on AI.
What Altman Actually Told the Council
Altman's central argument was that speed itself is a governance problem, not just a technical one: "As AI systems become more capable and more autonomous, they could move faster than our institutions, concentrate power in too few hands, or make decisions that people no longer understand or control." He was blunt about where the line sits: "The industry must not accept too much technological risk just because the benefits are too great and that they feel too important to slow down. Beating companies in a competitive race is not a reason to make rash decisions."
The line most outlets led with was his framing of who gets to decide: "If AI is to be democratic, the most important decisions cannot be made by labs in San Francisco alone. They must be shaped through democratic processes and by governments accountable to the people that they serve." He closed with an appeal to precedent: "In our history, there have been times where countries who compete and don't always like each other very much still come together for shared interests and the collective good in the face of a powerful new technology. We believe this must be one of those times." Concretely, he's pushing for shared benchmarks - standardized ways to measure both what a model can do and whether a company's safeguards actually hold up under testing.
What Amodei Actually Told the Council
Amodei's tone was more explicitly alarmed. His starkest line: "If managed poorly, I even believe that AI could be a risk to humanity as a whole." He named two specific dangers rather than speaking abstractly: bad actors using AI to help develop biological weapons, and model capabilities advancing past what developers can reliably control or verify.
His policy ask was concrete, not just a warning: a global ban on using AI to help develop biological weapons, new evaluation and verification systems, shared standards for testing models specifically for loss-of-control risk, and a notification system for flagging global security concerns as they emerge. He framed the stakes plainly: managing this "is bigger than any one company and has an industry-wide and global scale."
Bengio's Line That Cut Through Everything
Yoshua Bengio didn't soften anything. "The dangers are real and imminent," he told the Council, adding that this "unprecedented threat" was one "none of its members would choose." His most pointed moment was a direct rejection of the "we're in an unstoppable race" framing that's dominated policy discussion in Washington: "Citizens and governments should not accept the companies' claim that they are trapped in a race. The race is not a law of nature; it is the product of choices, choices made by the companies themselves. They've told us they would slow down if they could. They can."
The Part That Makes This More Than a War of Words
Here's what separates this from a normal policy disagreement. Fortune's reporting indicates Trump has already gone further than rhetoric: in a Truth Social post, he named Dario Amodei specifically and suggested his administration already holds "tremendous CRIMINAL and REGULATORY power" over AI companies. Separately, the Pentagon has reportedly designated Anthropic a supply-chain risk and stalled the release of its Mythos model - meaning there's an actual, material consequence already attached to Anthropic's public safety advocacy, not just a war of competing UN speeches.
That context reframes the whole exchange. Amodei isn't just making an abstract policy argument from a position of comfort - he's doing it while his own company appears to be facing direct government pressure over the exact stance he's publicly taking. Whether that's coincidence or cause-and-effect isn't confirmed, but the timing is hard to ignore.
Why This Matters Beyond the Speeches
The most important thing the session actually demonstrated: the AI industry is not unified behind the US government's deregulatory posture. OpenAI and Anthropic - genuine rivals who agree on very little day to day - stood in the same room (one virtually) and made essentially the same core argument, directly contradicting their own government's position from the previous day. That's a real split between the companies building the most advanced systems and the administration overseeing them, playing out in public, on the record, within a 24-hour window.
No binding agreement came out of the session - UN Security Council briefings don't produce enforceable rules on their own. What it did produce is a clear, documented record of where each side stands, right as the broader fight over AI regulation - covered in our AI Brake Crisis piece and antitrust lawsuit breakdown - keeps escalating on multiple fronts at once.
FAQ
What did Sam Altman actually say at the UN Security Council?
He argued AI is advancing fast enough to risk moving past what institutions and individuals can meaningfully follow or intervene in, said competitive pressure isn't a valid reason to cut corners on safety, and called for AI's biggest decisions to be shaped by governments and democratic processes rather than by labs alone.
What did Dario Amodei say?
He said poorly managed AI "could be a risk to humanity as a whole," named bioweapon misuse and loss of model control as the two central dangers, and called for a global ban on AI-assisted bioweapon development plus shared international standards for testing models.
How does this contradict Trump's UN speech?
Trump told the General Assembly the day before that AI safety warnings were a "hoax" and that the US "totally rejects" any coordinated global effort to control AI. Altman and Amodei argued the opposite: that global cooperation on safety standards is necessary and that companies alone shouldn't make AI's most consequential decisions.
Is Anthropic actually facing government pressure over this?
According to Fortune's reporting, yes - Trump named Amodei directly in a Truth Social post referencing his administration's "criminal and regulatory power," and the Pentagon has reportedly designated Anthropic a supply-chain risk, stalling its Mythos model's release. This detail comes from a single report rather than being independently confirmed across multiple outlets, so treat it as reported rather than fully settled.
What did Yoshua Bengio argue?
That AI companies aren't trapped in an unavoidable competitive race - the race is a choice they've made and can un-make - and that frontier AI development poses dangers he described as "real and imminent," not hypothetical.
Did the meeting produce any binding outcome?
No. UN Security Council briefings don't create enforceable rules on their own. The value was in establishing a clear, public record of where the leading AI labs stand, directly contradicting the US government's position from one day earlier.
Related Reading on AI Tech Safar
- Trump Renames AI to "Super Intelligence": The Full Story, Expert Reaction & What It Actually Means
- The AI Brake Crisis: Why the World's Biggest AI Labs Just Slammed on the Brakes
- AI Giants Sued for Collusion: Inside the Antitrust Lawsuit That Could Reshape the Industry
- Trump Announces "AI Force" and a New AI Czar - What's Actually Behind It
Useful Sources
- Bloomberg - Altman, Amodei Call on UN, World Leaders to Boost AI Safety
- Axios - AI leaders brief UN leaders on need for global safeguards
- AP via BNN Bloomberg - AI leaders warn UN of security risks as systems grow more powerful
- Fortune - Trump calls AI oversight a "globalist scheme" as Amodei and Altman head to the UN to ask for it
- CNBC - Altman and Amodei expected to join UN Security Council meeting about AI
- The Next Web - Yoshua Bengio urges UN Security Council to license frontier AI like nuclear power
تعليقات
إرسال تعليق