The U.S. is open to discussing shared risks with China in upcoming AI talks this weekend, Treasury Secretary Scott Bessent told Axios.
The Trump administration has flatly rejected calls from industry to pace or slow down the development of AI, citing a fierce race against China.
Bessent's scheduled meeting with Chinese Vice Premier He Lifeng could provide an opening for talks on U.S.-China AI safety cooperation.
"The United States remains the leader in AI, and we are open to discussions on avoiding shared risks and avoiding bifurcation of our two systems. We expect our discussions to cover both open and closed weight models," Bessent said in a statement to Axios.
"Bessent and [United States Trade Representative Jamieson Greer] are the point guys on China. So it makes sense the secretary is in this position," a senior administration official said.
There will be multi-hour meetings in New York City, focused on AI, trade, rare earths, and other economic issues, building on previous discussions held in cities around the world, sources familiar said.
The agenda will remain flexible to allow for substantive discussion across the range of relevant issues, sources added.
Possible working-level meetings will continue into Monday, sources said.
Bessent has played a key role in the Trump administration's AI agenda.
He was among the few cabinet members who flanked President Trump during a G7 press conference in June, where top AI CEOs and the administration discussed a U.S.-led effort to coordinate global AI standards.
Bessent also helped cool Anthropic's relations with the administration following a fallout with the Pentagon and is viewed as a reasonable actor who is trusted by the private sector and critical infrastructure operators, sources familiar have previously told Axios.
Commerce Secretary Howard Lutnick led high-profile AI discussions during the G20 summit in North Carolina earlier this month, while Bessent is taking the lead on talks with China.
Bessent's meetings in New York come days before Trump and Chinese President Xi Jinping's summit in Washington on Sept. 24.
AI CEOs have been calling for U.S.-China coordination on safety standards.
"All parties should jointly promote the open and inclusive development of AI for good and for all," the Chinese Foreign Ministry said this week, while pushing back on U.S. AI companies' calls to slow down and warning against "fear-mongering."
While a major breakthrough in U.S.-China safety cooperation remains a tall order, Bessent's involvement signals room for engagement on AI between the competing countries.
Posted by John3262005
2 Comments
Guess we will have to see if anything comes out this weekend
Big news maybe, I don’t know
My guess is that very basic KYC regulations will be implemented. Worse case is frontier models being legislated to require verified enterprise identity, phone-number risk scoring, corporate domain vetting, and corporate credit cards. This will have almost no impact on open models or models not at the frontier.
[https://www.euronews.com/2026/09/16/russia-used-claude-ai-for-espionage-disinformation-and-drone-swarms](https://www.euronews.com/2026/09/16/russia-used-claude-ai-for-espionage-disinformation-and-drone-swarms)
Regardless, the corporate risk Anthropic is assuming by seemingly allowing sanctioned actors to access frontier models to create drone swarms that kill Europeans is insane from a corporate risk management standpoint. It’s sheer luck that drone failed in Germany and we avoided a Claude drone mass casualty event.