UN Security Council Will Get Advice on AI Risks From Tech Giants Building It

The UN Safety Council will spend Wednesday afternoon listening to the individuals who construct the expertise it is most fearful about.

OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, AI security researcher Yoshua Bengio, and Hugging Face CEO Clément Delangue are all scheduled to temporary the 15-member council—the UN physique with the ability to authorize sanctions and army motion—on the dangers posed by synthetic intelligence.

DeepSeek is invited, too. The Chinese language startup that wiped roughly $600 billion off Nvidia’s market cap when it launched a cut-price mannequin in January 2025 has been requested to make an announcement, alongside one other Chinese language agency, Moonshot, builders of the Kimi fashions..

France organized Wednesday’s session and holds the council’s rotating presidency this month. Overseas Minister Jean-Noël Barrot will chair it, and the idea word behind the assembly reportedly facilities on AI methods slipping past human management.

The identical briefing is anticipated to the touch on AI that might ultimately run cyberattacks, sway elections, or design weapons with out a individual within the loop, which is why a council constructed for wars and sanctions is listening to from tech CEOs as an alternative of generals this week.

Amodei will be part of remotely somewhat than present up in individual. On September 12, he revealed an essay referred to as “We Should Tempo the Frontier,” arguing the business ought to intentionally sluggish how briskly it builds extra succesful frontier fashions—the most important, most superior AI methods any lab has launched. He stated Anthropic would give exterior evaluators everlasting, staff-level entry to test its security guarantees. Sam Altman and Elon Musk each stated they agreed inside hours.

Amodei referred to as China “the autocratic nation with by far probably the most superior AI capabilities” and argued the U.S. ought to hold limiting exports of superior chips to Beijing, even whereas asking Beijing to cooperate on international security guidelines. China’s Overseas Ministry referred to as the essay counterproductive, and state media branded it a “Chilly Struggle playbook” geared toward containing China.

President Donald Trump, in the meantime, desires no a part of the slowdown push. After Amodei’s essay dropped, the president referred to as AI-takeover warnings a “hoax” tied to the “Radical Left” and accused Amodei directly of wrongdoing with out providing proof.

Days later, King Charles hosted executives from Anthropic, OpenAI, Nvidia, and Google DeepMind at his Scottish property for an AI safety summit that produced dialog however no publicly recognized agreements.

Wednesday’s session is not more likely to produce a binding deal both. It lands two days earlier than Trump sits down with Chinese language President Xi Jinping in Washington on September 24, a summit the place AI security is reportedly on the agenda alongside commerce.