Sam Altman told the UN Security Council that no level of catastrophic risk from AI is acceptable.
Altman appeared as the second briefer at a Security Council session in New York today, in a meeting convened by France. He used the appearance to call for national and international standards governing frontier AI systems, rather than leaving the industry to police itself. The pitch was broad: establish a baseline for safety before these systems get more capable. It is an unusual venue for an AI company executive, since the Security Council typically handles wars and sanctions, not model weights.
Altman asking governments to write rules for the technology his company builds is not a new move for him, and AI executives making similar appeals is a familiar pattern by now. What is notable here is the venue. France convening a Security Council session on frontier AI signals that some governments are starting to treat model safety as a matter of international security, not just domestic tech policy.
Calling for regulation while continuing to ship products as fast as possible is a well-worn move in this industry. Nothing in what Altman told the Council suggests OpenAI's own pace is about to change.