A first for the Council
The UN Security Council meets in New York on Wednesday for a high-level briefing on artificial intelligence and international security, convened by France, which holds the Council presidency for September. The session is chaired by the French Minister for Europe and Foreign Affairs, Jean-Noël Barrot, and is being held during the high-level segment of the 81st session of the General Assembly, according to Security Council Report, which publishes the Council’s forward programme.
The Council has taken up artificial intelligence before. Security Council Report counts six earlier formal and informal meetings on the technology’s effect on international peace and security. Wednesday’s is the first devoted specifically to safety risks and to the possibility that humans lose control of advanced systems.
Who is briefing
The expected briefers are Yoshua Bengio, co-chair of the UN’s Independent International Scientific Panel on AI; OpenAI chief executive Sam Altman; Anthropic chief executive Dario Amodei; and Hugging Face chief executive Clément Delangue. It is an unusual line-up for a chamber whose witness list normally runs to special envoys and humanitarian coordinators.
Chinese developers were invited to speak as well. DeepSeek and Moonshot AI both received invitations to make statements, Runtime Wire reported on 22 September, adding that DeepSeek’s attendance, its delegate and its remarks were all still unconfirmed at the time of writing. Chinese state media has described Western calls to slow frontier development as an attempt to keep China out of the governance conversation.
What France put on the agenda
France circulated a concept note ahead of the meeting that emphasises safe and responsible AI development and raises systemic risks from misalignment and loss of control over advanced models. The note singles out the prospect of autonomous systems attacking critical infrastructure, and argues that artificial general intelligence capable of recursive self-improvement raises the loss-of-control risk further.
It sets four questions for members: what risks to international peace and security follow from the pace of AI progress; how governmental and non-governmental actors can address them; which diplomatic tools, coordination measures and confidence-building steps are available; and whether evaluation and verification methods can keep pace with the systems they are meant to measure.
No outcome document expected
Security Council Report does not expect a formal agreement. Members are likely to concur that safeguards are necessary while disagreeing about who holds regulatory authority, how far restrictions should go and how anything agreed would be enforced — the familiar split between national and international governance frameworks.
The briefing follows a thematic brief published by the UN’s Independent International Scientific Panel on AI on 21 September, which argued that the traditional model of safeguarding is coming apart as systems act with less human supervision. Bengio co-chairs that panel and will carry its findings into the chamber.
What to watch
Two things. The first is whether DeepSeek speaks at all, and what a Chinese frontier developer says in front of the Council when the subject is losing control of its own products. The second is whether the Council treats this as a one-off briefing or the start of a work stream — a presidential statement or a resolution would be the signal, and neither is on the table this week.
