OpenAI and Anthropic CEOs Tell UN Security Council AI Requires Global Regulation
AI industry leaders warned the United Nations Security Council that artificial intelligence could become an uncontainable global threat if left without coordinated oversight.
The CEOs of the two leading frontier AI companies told the United Nations Security Council this week that their industry represents a threat so vast it demands international coordination to contain it. OpenAI’s Sam Altman and Anthropic’s Dario Amodei addressed the UN Security Council on AI safety, delivering warnings that have become a familiar refrain from their sector but which carried new weight in the chamber responsible for maintaining global peace and security.
Sam Altman warned humanity could ‘lose control of the future of AI’ during the session, framing the current moment as a narrow window for establishing governance before the technology becomes uncontrollable. Altman advocated for what he described as democratic oversight of the technology’s development. Dario Amodei echoed the existential tenor of the warnings, stating ‘If managed poorly, I even believe AI could be a risk to humanity as a whole’.
The briefing, which included remarks from Turing Award winner Yoshua Bengio, presented the Security Council with a stark picture of both huge benefits and huge risks. Bengio framed advanced artificial intelligence systems that operate beyond human control as an unprecedented class of threat. He told the Council that such a threat represents something that none can contain alone and that does not respect the borders we defend.
This appeal for global regulation from the leaders of the very firms building the most advanced models marks a continuation of a lobbying campaign that has sought to shape oversight before it is imposed upon them. The choice of venue, however, underscores the scale of the perceived risk. The Security Council typically convenes to address interstate conflict, weapons proliferation, and threats to international peace—matters where unilateral national action is understood to be insufficient. The framing of frontier AI as a borderless, existential risk that could threaten the whole world is an explicit argument for treating the technology in that same category.
The heads of several major AI firms told the UN Security Council their industry urgently needed global oversight to avoid dangers that could threaten the whole world. The session itself was characterized by the UN as addressing a ‘real and imminent’ threat posed by runaway AI, language that borrows directly from the lexicon of security briefings on nuclear or biological weapons.
What remains less clear is the shape of the governance the executives are advocating for, or the mechanisms by which a fractured international community could achieve it. Altman’s call for democratic governance and Amodei’s repeated warnings create a consensus on the problem but not a roadmap for the solution. The Security Council, for its part, has no established regulatory apparatus for technology; its power lies in issuing binding resolutions, imposing sanctions, and authorizing the use of force—tools poorly suited to governing the development of software.
The briefing thus leaves a central tension unresolved. The industry’s most influential figures are now explicitly asking the world’s highest security body to treat their product as a potential weapon of mass disruption, while the body itself lacks any obvious technical or legal framework to do so. The warnings have been delivered to the room most accustomed to contemplating civilization-level risks. The question of what that room can actually build, and whether the companies would accept its authority once constructed, is the next, much harder chapter.