President Donald Trump, in his speech to the United Nations General Assembly on Tuesday, rejected the idea of international oversight of artificial intelligence—what he has just taken to calling “superintelligence.” In thus address, he called such proposals a “globalist scheme” to control the technology.
A day later, the heads of America’s two leading AI labs will ask the UN for close to the opposite.
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei are on the schedule to brief the UN Security Council Wednesday afternoon at a session on AI and international security. The session will be livestreamed.
Altman is expected to discuss the company’s safety work, and the importance of international cooperation, an OpenAI spokesperson told Reuters.
Trump also took direct aim at AI safety advocates, who have been bashed by both the MAGA crowd and by venture capitalists suspicious of effective altruism, the once-niche philosophy that takes a utilitarian perspective on risk. He drew a comparison belied by a rapidly boiling atmosphere: the same people who warned about global warming, he said, “are now saying that AI is going to kill us all.”
He signaled that he would not listen to those calls. “I’m not going to stifle growth of something that will be bigger than the industrial revolution,” Trump said.
The industry’s own alarm
Trump’s stance puts him at odds with the top executives at the labs building the technology. On Sept. 12, Amodei published an essay asking AI companies to slow the pace at which they improve their models and asked for independent evaluators to get complete access to frontier labs. Altman and Elon Musk publicly supported the call, though Nvidia CEO Jensen Huang called the fears overblown.
After Amodei’s essay, Trump rejected those fears on Truth Social, calling the existential threat a “hoax” and that the only safeguard AI needed was a “high IQ” president. In the same psot, he seemed to hint at penalties against AI companies, saying that his administration already had tremendous CRIMINAL and REGULATORY power” over them, and he singled out Amodei by name.
Indeed, the administration has wielded its power against Anthropic, with the Pentagon previously designating the company a rare supply-chain risk and stalled the release of its Mythos model.
Those warnings followed a string of incidents; OpenAI disclosed in July that its AI models broke out of its “sandboxed” testing environment and autonomously hacked into the open-source developer platform Hugging Face. Anthropic and Google Gemini have also disclosed safety incidents.
Hugging Face CEO Clément Delangue will also brief the Council on Wednesday, alongside Yoshua Bengio, co-chair of the UN’s Independent International Scientific Panel on AI. France, organized the session to address malicious use of AI and the risk of losing control of advanced models. Chinese developers DeepSeek and Moonshot were also invited to make statements, though it is unclear if they will attend.
A world split on rules
Ahead of the General Assembly, leaders of more than 20 mostly European countries urged binding safety measures for advanced AI, warning the technology could outpace the ability to control it. The two leaders of the technology, the U.S. and China, both abstained from supporting the idea.
Trump is set to meet Chinese President Xi Jinping, and the two are expected to discuss AI. Trump will also host a state dinner for Xi, with Nvidia CEO Jensen Huang, Altman, and Apple’s Tim Cook expected to attend.