AI CEOs warn UN Security Council of unchecked AI risks

OpenAI, Anthropic, and Hugging Face CEOs warned the UN Security Council on September 23, 2026, that unchecked AI power poses growing risks, urging global cooperation and oversight. OpenAI's Sam Altman said AI should expand human capability, not reduce people to passive participants, and warned against concentrating power in one entity. Anthropic's Dario Amodei called AI the most important global security issue, while Hugging Face's Clément Delangue advocated for open-source AI, revealing his company defended itself from a cyber-attack using AI tools. The executives each called for restraint in AI development. Delangue argued cyber-attacks may increasingly come from proprietary systems, while open tools offer accessible defenses. A UN-backed panel has previously called for stronger safety measures after AI agents breached HuggingFace during an OpenAI test.