Treasury Secretary Scott Bessent said on Sunday the US has proposed a new “notification mechanism” for artificial intelligence incidents that could impact national security. This was part of weekend discussions ahead of talks between President Donald Trump and China’s Xi Jinping at the White House this week.
“We want a common vision of common goals and common threats,” Bessent told reporters after talks with Chinese Vice Premier He Lifeng in New York. “We believe that, as with any cross-border activity, it is very important to move from opacity to greater transparency between the world’s No. 1 and 2 AI powers.”
Trump resisted calls to slow AI development, saying it would help China catch up with U.S. companies.
As Bessent and U.S. Trade Representative Jamieson Greer stood together late Sunday in the towering lobby of JPMorgan Chase headquarters, they said the groundwork had been laid for talks between Xi and Trump scheduled to take place on Thursday. They said the two sides had agreed to set up a new “chamber of commerce,” which the two leaders discussed when they met in Beijing in May.
“Today, those ideas have become reality,” Greer said.
Greer said countries were still discussing which “insensitive” items could be considered for lower tariffs, but the list could potentially include consumer and agricultural goods, energy products and medical devices.
Chinese state media Xinhua said Bessent and the vice premier discussed topics “related to AI” without giving details, and also had “candid, detailed and constructive” discussions on key economic and trade issues.
“The fact that both sides have agreed to continue dialogue is significant,” George Chen, partner and head of the digital practice at consultancy The Asia Group, said in written comments. “The first outcome – an AI risk notification mechanism – sets a precedent that other countries could follow.”
Related: What security leaders are learning from the Hugging Face incident about AI agent access
Related: Industry Reactions to OpenAI Models Hacking Hugging Face
Related: OpenAI agents are coordinated through a makeshift message board before the Hugging Face Hack
