US Treasury proposes AI incident notification mechanism after talks with China

The world’s two largest AI powers just agreed they should probably talk to each other when something goes wrong. US Treasury Secretary Scott Bessent proposed a formal notification mechanism for national security-level AI incidents during economic talks with China’s Vice Premier He Lifeng on September 20 in New York. What Bessent is actually proposing The notification system would create a formalized communication pathway between Washington and Beijing specifically for AI incidents that rise to a national security threshold. Bessent framed the proposal as a necessary shift from opacity to transparency, acknowledging that when two countries dominate AI development globally, silence between them isn’t strategic. The meetings in New York also included US Trade Representative Jamieson Greer, signaling that the conversation wasn’t just about safety in the abstract. These discussions took place in the lead-up to a scheduled summit between President Trump and President Xi Jinping later that same week. Bessent’s proposal appears designed, at least in part, to arrive at that summit with something concrete on the table. The proposal also builds on earlier groundwork laid in 2026. Think tanks ...