OpenAI, Anthropic CEOs call for global AI regulation at UN

Al Jazeera reports that the heads of several major AI companies addressed the United Nations Security Council on Wednesday, saying the sector urgently requires international supervision because of risks that could endanger everyone. The session was convened by France and coincided with the General Assembly gathering in New York, at a time when experts increasingly warn that fast AI development needs more human supervision.

Anthropic chief executive Dario Amodei said that, if handled poorly, AI could endanger humanity overall. OpenAI's Sam Altman warned the 15-member body that people might lose their grip on where AI takes the world, and argued that the most important decisions should not be made by San Francisco labs alone but through democratic processes and accountable governments. Several council members, including the French and British foreign ministers, agreed that the international community should create shared frameworks. Hugging Face CEO Clement Delangue said his company, attacked by out-of-control models, used a Chinese model to defend itself, since that model was subject to fewer limits than US equivalents. Yoshua Bengio, who co-chairs the UN's independent scientific panel on AI, called the dangers real and imminent.

The article stresses reluctance in Washington and Beijing. The Trump administration's representative, Michael Kratsios, said Washington rejects, in full, any attempt by international bodies to impose central control or worldwide governance over AI. Experts say the technology race between the US and China makes new restrictions less likely. Daniel Forti of the International Crisis Group said the UN is coming to recognise the danger and the need for some rules.

For background, Al Jazeera notes that the General Assembly passed a nonbinding AI resolution in 2024, and created a scientific panel and the Global Dialogue on AI Governance last year.