Altman and Amodei argue separately before the UN Security Council for international rules on AI

The two rivals spoke on 23 September in New York. Amodei sees it as "the main global security issue", Altman calls for decisions by democratic institutions.

Sam Altman (OpenAI) and Dario Amodei (Anthropic) spoke on September 23, 2026 before the United Nations Security Council. Head-on competitors in the market, the two leaders spoke separately but agreed on one point: without international standards or cooperation between states, the current trajectory of AI becomes a collective security problem.

Amodei put three avenues on the table: narrowly targeted global agreements, starting with a ban on the use of AI to design biological weapons; assessment and verification mechanisms for states to monitor each other's compliance with commitments; and common standards for testing models, with a mechanism for reporting security incidents. He called the issue "the main global security challenge facing the world today" and warned that mismanagement would put all of humanity at risk.

Altman, for his part, insisted that structuring decisions should be made by democratic institutions "accountable to the people they serve" — a formulation that effectively excludes some countries that are very advanced in AI, China in the lead. The session was a briefing and not an adoption session: no new rules came out of it.