AI chiefs asked the UN for regulation — the US said no to global governance
When the UN Security Council met on artificial intelligence on 23 September 2026, the chief executives of OpenAI and Anthropic asked to be regulated — warning that AI could become a risk to humanity itself. But the meeting exposed a hard divide among states: the White House rejected any new global governance structure, while China called for more regulatory cooperation — with no confirmed formal statement or follow-up from the meeting.
Warnings from the industry's own leaders
The Security Council meeting took place on Wednesday 23 September 2026, convened by France during the UN General Assembly's high-level week, according to Associated Press via the Baltimore Sun. OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei addressed the council, urging it to regulate the very technology they build.
Altman said, according to AP: "We could lose control of the future to AI." He also said, according to ABC News, that "this moment calls for extreme care." "If AI is going 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 they serve," Altman said, according to ABC News.
Amodei went further. "If managed poorly, I believe AI could even be a risk to humanity as a whole," he said, according to AP. Both CEOs told the council that their companies commit to slowing releases of new models if they recognize a threat, ABC News reports.
Yoshua Bengio — one of the researchers described as a "godfather of AI" and co-chair of the UN's independent international scientific panel on AI — warned that the risks of advanced AI are no longer theoretical, according to Newsweek.
The day before the meeting, UN Secretary-General António Guterres had warned about "killer robots" in the General Assembly. Experts and European leaders said, according to AP, that this was no exaggeration, but something that is either already a gruesome reality or dangerously close.
The host framed the stakes. French Foreign Minister Jean-Noël Barrot compared today's AI to both the runaway computer HAL9000 in the film "2001: A Space Odyssey" and the dawn of the atomic age, according to AP.
The states divided
The meeting exposed a clear political divide among member states.
The US took the sharpest distance from global regulation. Michael Kratsios, the White House's science and technology adviser, said that concerns about controlling AI are "not a reason to halt further development or restrict it with a new global governance structure," according to AP. Countries should instead build their own expertise and exchange effective practices rather than create a worldwide regulatory regime, he argued: "You cannot govern technology you do not understand," Kratsios told the council, according to Firstpost.
China took the opposite position. The country's UN ambassador, Fu Cong, called for further development of regulatory systems, stronger emergency response and closer cooperation between countries, according to Firstpost.
The New York Times describes a council divided over which global AI measures should be enforced at all, with the US and China showing little interest in slowing development — amid deep mutual suspicion between them.
The UK sided with the CEOs. Foreign Minister Ed Miliband said, according to AP, that the tech leaders are "telling us in plain terms that it cannot be left to them to prevent catastrophe from hitting the world," adding: "We cannot ignore them, because they are right." According to AP, he pledged to put AI control at the heart of the UK's G20 presidency.
Words and action
ABC News points to a tension in the industry's own position: both CEOs told the council that their companies are willing to slow the release of new models if they recognize a threat. Yet in the week after the CEOs originally called to "pace the frontier" of AI development, both companies have released new models. Whether the releases conflicted with the companies' own thresholds or were compatible with the pledges has not been clarified in the available material.
The same day the Security Council met, Senator Bernie Sanders and Representative Greg Casar — both progressives — introduced legislation that would ban artificial superintelligence and create a federal agency to oversee advanced AI, according to AP (via the Boston Herald). The proposal came as parts of the industry itself urge Congress to introduce safeguards; its prospects of passage are not addressed in the material.
During the debate, Hugging Face CEO Clément Delangue gave, according to Al Jazeera, his own account of the company being attacked by AI and defending itself with AI — including a Chinese model with fewer restrictions than comparable American tools. This account comes from Delangue himself and is not independently confirmed in detail.
What remains unresolved
There is no confirmed formal statement, resolution or follow-up commitment from the meeting. Several caveats also apply to the sources: all quotes are relayed through news organizations, and there are minor discrepancies in wording between them — for example, Al Jazeera renders Altman's warning as "the future of AI," while AP has "the future to AI." Such discrepancies cannot be cross-checked against an official transcript.
- Feasibility: Neither China's call for regulation nor the UK's G20 pledge is binding, and the sources do not describe how any of the measures would be implemented.
- The US–China gulf: With a White House rejecting global governance, a China demanding more cooperation — and both countries holding vetoes on the council — no path to agreement is visible in the material.
- The right forum: Kratsios' objection was precisely that countries should focus on their own expertise and exchange of practices rather than a global regulatory regime — a principled question the meeting did not answer.
What was nonetheless remarkable on 23 September was the shape of the debate: the most central leaders of the AI industry themselves appeared before the Security Council and asked to be regulated — while the most powerful state in the room wanted it differently.