AI walks into the UN Security Council: what today's briefing means for security teams
On September 23, 2026, the chief executives of OpenAI and Anthropic addressed the United Nations Security Council on AI safety and international security risks. Sam Altman joined in person; Dario Amodei participated remotely. They were joined by Hugging Face CEO Clement Delangue and Yoshua Bengio, co-chair of the UN's Independent International Scientific Panel on AI. DeepSeek and Moonshot, two Chinese AI labs, also made statements. France, holding the September Security Council presidency, convened the 15-member session.
This was the first time AI lab leadership has addressed the UN Security Council directly on AI risk. That alone makes it a signal worth tracking.
What Amodei said
Amodei reiterated his call for slower development of the most powerful AI systems, a position he has held publicly since early 2026. He is also backing continued US semiconductor export restrictions on China, a significant statement to make inside a chamber where China holds a permanent seat with veto power. Altman and Elon Musk of xAI have backed Amodei on the slowdown argument. Whether that consensus holds as competitive pressures intensify is a different question.
What this means for security teams
The UNSC session matters less for what it decides and more for what it signals about where the regulatory conversation is heading. Three observations for security practitioners.
First, AI is now on the agenda of institutions that set binding international law. The UNSC cannot regulate AI directly, but a Security Council resolution on AI would carry significant diplomatic weight. Security and compliance teams in multinational organizations should expect AI-related disclosure and governance requirements to follow a similar curve to data protection regulation after GDPR: slow adoption, then fast.
Second, the joint appearance of OpenAI, Anthropic, and DeepSeek at the same table is a quiet acknowledgment that AI risk is not a Western problem. Chinese AI labs participating in a UNSC session is a meaningful change in tone. It does not resolve the competitive dynamics, but it creates a precedent for multilateral AI safety dialogue that includes the labs most capable of producing frontier-level risk.
Third, Amodei's call for slower development at the UNSC is notable because it aligns the CEO of Anthropic publicly with a regulatory position that constrains his own company's competitive position. That is an unusual move, and it suggests Anthropic genuinely believes the risk is real enough to accept the cost of saying so in a formal international setting.
Gigia Tsiklauri is a Security Architect and founder of Infosec.ge. Get in touch if you want to discuss AI governance implications for your organization.