Senior American and Chinese officials have concluded roughly eight hours of talks in New York, with Washington proposing a formal communication channel for sharing information about artificial intelligence threats — marking a rare moment of potential cooperation between the world's two leading AI powers. US Treasury Secretary Scott Bessent, who met Chinese Vice Premier He Lifeng at JPMorgan Chase's Manhattan headquarters on Sunday, described the discussions as "very successful" and said both sides had agreed to further working-level meetings.
The centrepiece of Washington's AI proposal is a so-called "US-China AI dialogue" that would include a notification mechanism — a system allowing either country to alert the other about AI incidents serious enough to raise national security concerns. Bessent framed the initiative in terms of transparency, arguing that moving "from opaque to more transparency between the number one and the number two AI powers in the world is very important." US Trade Representative Jamieson Greer, who also participated in the talks alongside China's chief trade negotiator Li Chenggang, stressed that export controls on advanced AI chips and semiconductor manufacturing equipment would remain outside the scope of any such arrangement. Chinese state media described the meetings as "candid, in-depth, and constructive," though Beijing's delegation did not address reporters directly.
The talks come at a moment of heightened concern about advanced AI's risks, even as China and the United States remain locked in fierce technological competition. Analysts note that frontier US models still lead their Chinese counterparts, but the gap has been narrowing. Beijing this week released the third iteration of its AI safety governance framework, addressing risks from autonomous AI agents, cybersecurity threats and the theoretical possibility of AI systems improving themselves recursively. While China's foreign ministry has pushed back against Western calls for a slowdown — characterising them as an attempt to lock in American advantage — researchers say there is more common ground on underlying safety concerns than the rhetoric suggests. "In many ways, I think there is more common ground than people might assume," said Gabriel Wagner of Concordia, a Beijing-based AI safety consultancy.
The AI discussions formed part of broader preparatory talks ahead of a planned White House summit between President Donald Trump and Chinese President Xi Jinping later this week. The two leaders are also expected to meet in November at the Asia-Pacific Economic Cooperation forum in Shenzhen, and potentially again at the G20 in December. Beyond AI, Sunday's talks touched on trade: officials discussed identifying "non-sensitive" goods — potentially Chinese consumer products and American energy, agricultural goods and medical devices — that could qualify for lower tariffs. A fragile trade truce, reached in Busan, South Korea, last October and capping US tariffs on Chinese goods at around 20 percent, is approaching its 12 November expiry. Analysts cautioned against high expectations for the summit, with former US Deputy Secretary of State Kurt Campbell saying the most realistic near-term outcome is "a truce" that holds through the end of the year.