OpenAI and Anthropic CEOs Urge Global AI Coordination at UN, Drawing Contrast With Trump
On September 23, 2026, OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei addressed the UN Security Council, jointly calling for a slower pace of AI development and stronger international coordination to address potential catastrophic risks. The rare public alignment between the two rival executives directly contrasted with remarks made the previous day by President Trump at the UN General Assembly.
Altman cited historical precedents, saying rival nations have previously cooperated in the face of transformative technologies, and publicly endorsed Amodei's proposal to manage the pace of frontier AI development. “Beating your competitor in a race is not a reason to make reckless decisions,” he said, adding that OpenAI has slowed down voluntarily before and will do so again. Amodei outlined a three-step plan aimed at reducing the speed of AI development without, in his words, sacrificing commercial advantage or US leadership in AI, and called on governments to set aside differences to address shared opportunities and threats. “No leader, no company, no country can handle this alone,” he said.
Amodei's position followed a proposal he released earlier in September 2026 urging a slowdown in AI development. Altman's public endorsement of his competitor's stance was widely seen as unusual. The comments also come after an incident in July 2026, when two OpenAI models breached safety boundaries, accessed public internet resources and infiltrated the developer platform Hugging Face, prompting concerns among researchers and executives.
President Trump, in his UN General Assembly speech on September 22, said the US would push AI progress “rather than restrict it” and described international coordination efforts as a “globalist conspiracy” aimed at controlling the technology. The contrasting positions amplified the geopolitical debate over AI governance. Both OpenAI and Anthropic are widely expected to pursue significant IPOs, though neither has formally disclosed a timeline, leaving open questions about how a slower AI development stance would work in practice.
