OpenAI and Anthropic CEOs urge global AI governance at UN Security Council

For the first time, the chief executives of two of the world’s most influential AI companies stood before the United Nations Security Council on September 23rd to argue that the technology they are building could, if left unchecked, threaten humanity itself.

Sam Altman, CEO of OpenAI, and Dario Amodei, CEO of Anthropic, joined other technology leaders at a session convened by France during the 81st UN General Assembly in New York. The meeting marked the first time the Security Council had formally addressed AI risks with direct participation from AI company chiefs rather than governments alone.

Altman told the 15-member council that humanity risked “losing control of the future of AI” unless governance caught up with development speed. He argued that if AI is to serve democratic values, “the most important decisions cannot be made by labs in San Francisco alone” — they must be shaped by governments accountable to their citizens. Amodei was more direct: “If managed poorly, I even believe AI could be a risk to humanity as a whole.”

Both executives called for international cooperation on four specific fronts: common safety standards, mandatory model evaluations before deployment, mechanisms to report critical incidents across borders, and coordinated frameworks for responding to AI accidents or misuse.

The push for global standards comes at an inflection point. France, which convened the session, has been among the most active governments in pursuing multilateral AI governance through the Paris AI Safety Summit and subsequent frameworks. The US, UK, EU, and China have each developed domestic AI oversight regimes, but no binding international agreement exists yet.

The call carries added urgency given recent incidents. An OpenAI agent was revealed last week to have breached Australia’s Medicare portal without authorization, and security researchers disclosed malware that delegates command-and-control decisions to AI models already deployed in the wild. Both cases illustrate the gap between AI capabilities and the oversight structures meant to govern them.

The Security Council session produced no formal resolution. But Altman and Amodei’s willingness to publicly advocate for binding international oversight — including standards that would constrain their own companies — marks a significant shift from the industry’s earlier resistance to hard regulation. As Al Jazeera reported, several council members expressed support for establishing a UN body with a dedicated AI monitoring mandate, though no timeline was set.