U.S. and China propose AI risk notification mechanism ahead of Trump-Xi talks

The U.S. has put forward a new notification system for AI incidents that could affect national security, according to Treasury Secretary Scott Bessent after weekend talks with Chinese Vice Premier He Lifeng in New York. The pitch is simple enough: if something serious happens, the two biggest AI powers on the planet would have a way to tell each other about it instead of operating in the dark.

Bessent framed it as a move from opacity toward transparency, and he tied it directly to the White House talks scheduled for Thursday between President Donald Trump and China’s Xi Jinping. Standing with U.S. Trade Representative Jamieson Greer at JPMorgan Chase headquarters, he said the groundwork for those talks had already been laid.

The two sides also said they had agreed to make real a new “Board of Trade” idea that Trump and Xi discussed in Beijing in May. Greer put it bluntly: what had been an idea is now becoming reality.

Trade remains part of the picture, too. Greer said the countries are still talking about which “nonsensitive” items could qualify for lower tariffs, with possible candidates including consumer and agricultural goods, energy products and medical devices. Chinese state media Xinhua described the AI talks more vaguely, saying only that the sides discussed issues “relating to AI” and had “candid, in-depth and constructive” exchanges on economic and trade issues.

The most telling reaction may have come from George Chen of The Asia Group, who said the continuation of dialogue matters on its own. He argued that the AI risk notification idea could become a model for other countries, which is exactly the kind of small, practical move that can matter more than the grand speeches around it.