A notification channel for AI incidents that could threaten national security is now on the table between the world’s two largest AI powers, after a full day of negotiations in Manhattan.
Treasury Secretary Scott Bessent told reporters the aim is a shared understanding of common goals and common threats. Moving the leading and second-ranked AI nations from opaque dealings toward more transparency matters, he said, and negotiators will meet again to settle how the reporting system works.
Bessent spoke at JPMorgan’s headquarters after talks with Vice Premier He Lifeng. Both delegations also put into operation a board of trade sketched out during President Donald Trump’s May visit to Beijing. Advanced AI chip export controls never came up, US Trade Representative Jamieson Greer said.
Timing matters. The UN General Assembly is gathering, Trump hosts Xi Jinping at Joint Base Andrews on Wednesday, and the summit itself follows on Thursday. Trade negotiators are expected to weigh an extension of the Busan truce, with Washington wary of Beijing’s follow-through on rare earths.
Lab leaders had pushed for a wider slowdown agreement. Trump has refused, arguing the US holds a lead worth keeping, and Nvidia chief Jensen Huang told CBS News that AI companies should face existing cybersecurity and liability law first.
