NEW YORK — OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei briefed the United Nations Security Council on 23 September 2026, warning that the rapid development of large‑language models could pose a “real and imminent” threat to global security. The two AI firms presented a joint statement that called for a new international framework to regulate the deployment of advanced generative models.
UN Security Council meets AI leaders
The briefing took place in the Security Council chamber at the UN Headquarters, following a request from the council’s AI Governance Working Group. Altman and Amodei were accompanied by senior staff from their respective companies, including chief technology officers and policy advisors. The meeting was attended by the council’s nine permanent members and a rotating group of non‑permanent members, as well as representatives from the UN Office for the Coordination of Humanitarian Affairs and the UN Office of Counter‑Disinformation.
During the session, the CEOs outlined the technical capabilities of their latest models, noting that the models can generate realistic text, images, and audio with minimal human oversight. They highlighted the potential for misuse in disinformation campaigns, cyber‑attacks, and autonomous weapon systems. The briefing also included a demonstration of a model’s ability to produce persuasive political messaging, which the speakers said could undermine democratic processes if left unchecked.
Key points from the briefing
Both companies stressed that the current regulatory environment is insufficient to manage the risks associated with large‑scale AI systems. They urged the council to adopt a binding treaty that would set standards for model safety, transparency, and accountability. The joint statement also called for the establishment of an independent oversight body to monitor compliance with the treaty.
Altman and Amodei cited recent incidents in which AI‑generated content was used to manipulate public opinion during elections in several countries. They warned that the speed at which AI can produce and disseminate content far outpaces existing legal frameworks. The CEOs also highlighted the need for international cooperation in research and development to ensure that safety protocols are built into models from the outset.
UN response
The Security Council’s spokesperson, Ms. Maria L. Hernandez, said the council had received the briefing and would consider the proposals in its upcoming agenda. She noted that the council would consult with member states and relevant UN agencies before moving forward. The spokesperson added that the council had already been working on a draft resolution on AI governance, which would be tabled in the next session.
UN Secretary‑General Antonio Guterres, who was present at the briefing, said the council’s engagement with the AI industry was a positive step toward addressing emerging security challenges. He emphasized that the UN’s role is to facilitate dialogue and ensure that technological progress does not outpace the development of international norms.
Member states expressed a range of views. The United States and the United Kingdom welcomed the call for a treaty, while China and Russia urged caution, citing concerns about sovereignty and the potential for the treaty to be used to restrict technological innovation. The council agreed to form a sub‑committee to draft a framework that balances innovation with security.
