Altman takes his 2015 warning to the UN Security Council

In a 2015 essay titled "Machine Intelligence, part 1", Sam Altman wrote that the "development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity". Next Wednesday he brings that theme into the UN Security Council chamber: an OpenAI spokesperson told Reuters that Altman will brief an open Council meeting in person.

At a glance

• The 15-member Security Council is due to meet on Wednesday on artificial intelligence and international security, during the week world leaders gather in New York for the General Assembly.
• France holds the Council presidency for September, convened the session and circulated a concept note to members; French foreign minister Jean-Noël Barrot will chair the meeting.
• Diplomats also expect a high-level presence from Anthropic, though that was not confirmed, and nothing announced so far describes what the Council would do after the briefing itself.

If you did not follow the first round, the Council took up AI risks in 2023, when China said the technology should not become a "runaway horse" and the United States warned against its use to censor or repress people. According to Security Council Report, that session on 18 July 2023 was chaired by then-UK Foreign Secretary James Cleverly, with briefings from Secretary-General António Guterres, Anthropic co-founder Jack Clark and Yi Zeng of the Chinese Academy of Sciences.

France convened the session and Jean-Noël Barrot will chair it

The meeting falls on Wednesday, with the 15-member Council taking up artificial intelligence and international security. France, president of the body for September, called it, and French Minister for Europe and Foreign Affairs Jean-Noël Barrot chairs.

In a concept note circulated to Council members and seen by Reuters, France wrote:

In light of the risk of misuse of this technology for malicious purposes, and drawing on the Council's existing engagement in this field, France wishes to underline the urgency of action to foster safe and responsible development of artificial intelligence for the benefit of the international community as a whole.

The same note says the session should let participants "present their analysis of the risks linked to loss of control or the use of the latest AI models to develop, facilitate and conduct actions with an impact on international security and peace".

A rotating presidency decides what the Council talks about

Here is the machinery behind that sentence. The presidency of the Council rotates month by month, and whoever holds it shapes the programme of work, which is how AI lands on the agenda in September rather than at some other point.

A concept note is the preparation document: it goes round to members in advance and tells them what the discussion is for and which questions to come ready to answer. Think of it as the agenda a conference organiser sends before a panel, so nobody arrives improvising. An open meeting means the session is public rather than held behind closed doors, and briefers are outside voices invited to speak to the members.

Altman is expected to speak about shared safety standards

The OpenAI spokesperson said Altman's remarks are expected to focus on the need for international coordination and shared safety standards, along with the steps OpenAI is taking so that people globally benefit.

The timing sits close to something else: the briefing comes weeks after several AI industry leaders called for a coordinated slowdown of the development of the technology, warning that it could soon improve on its own and slip beyond human control. Altman's own warnings predate the current boom by years.

According to the Center for AI Safety, Altman is listed as a signatory, under "CEO, OpenAI", of its Statement on AI Extinction Risk, which reads: "Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war." The same list, per CAIS, carries Dario and Daniela Amodei, Geoffrey Hinton, Demis Hassabis and Ilya Sutskever.

Diplomats expect Anthropic at the table as well

Diplomats said they expected high-level presence at the meeting from Anthropic, but that it was yet to be confirmed. That matters for the misuse half of the French note, because Anthropic has been publishing what misuse looks like in practice.

In Anthropic's own reporting, its September 2026 report "Detecting and countering misuse of AI" covers malicious use of Claude models disrupted between December 2025 and August 2026, across cyber operations, influence operations, surveillance, biological misuse, conventional weapons development and illicit distillation, following earlier reports in March, August and November 2025. Anthropic labels the actors it observes with internal designators it calls Generative Threat Groups, and measures "uplift", the capability boost an attacker gets from the model, through speed, scale and depth.

What none of the announcements describe is an outcome. Neither the OpenAI spokesperson's account nor the French concept note points to a text, a resolution or a follow-up date, and the Anthropic attendance is still unconfirmed. In our view the gap worth watching is between a note that underlines the urgency of action and a format that is designed to collect analysis.

Whether Anthropic takes a seat

The open meeting is on Wednesday, in the middle of General Assembly week in New York, so the speaker list is the first thing to check: diplomats expect a senior Anthropic figure, and that has not been confirmed. Two other things are unknown from what France and OpenAI have said so far, namely whether any other labs are invited to brief, and what the Council does with the session once it ends.