US breaks with allies on AI safety rules at UN
Trump's science adviser called international AI safety efforts an "intervention," leaving the US isolated as France and China pushed for a shared framework to manage frontier-model risk, Semafor reported.
The Trump administration rejected calls for binding international AI safety rules at the United Nations General Assembly this week, Semafor reported Friday. White House science adviser Michael Kratsios called such efforts an "intervention." President Trump said AI development should stay "exactly where it is," prioritizing speed over new oversight.
China's Xi Jinping took the opposite position, backing a global AI governance framework, Semafor said. French official Jean-Noël Barrot chaired a UN meeting on AI and security this week. He said the international community needs "a common framework to manage the risks associated with frontier AI models." The gap left the US without close allies on the question.
The talks took place as Australian Prime Minister Anthony Albanese used the same UN gathering to press for tougher AI safeguards. He cited OpenAI's breach of a government Medicare portal, disclosed earlier this week. Nations appeared headed toward nonbinding rules and best practices rather than enforceable standards, Semafor said.
The divide was not confined between nations, Semafor reported. Some AI company chief executives and congressional Republicans have pushed for stronger safety measures, putting them at odds with their own administration. No binding commitments came out of the meetings, according to the report.