U.S. Proposes AI Safety Mechanism Before Trump-Xi Summit

Senior U.S. and Chinese officials met in New York on Sunday to prepare for this week's summit between President Donald Trump and Chinese President Xi Jinping, and the discussions included a U.S. proposal for a bilateral AI safety arrangement, the BBC reported. Treasury Secretary Scott Bessent, who met with Chinese Vice Premier He Lifeng, told reporters the two sides talked about creating a "notification mechanism" for AI incidents that could affect national security. He described the talks as successful.

Bessent framed the proposal as a step toward greater openness between the two countries. "Moving from opaque to more transparency between the number one and the number two AI powers in the world is very important," he said, according to the BBC. He added that the meeting also covered plans to operationalize a process, called the Board of Trade, for identifying potential tariff cuts on goods.

Chinese state news agency Xinhua said the two sides held candid, in-depth and constructive exchanges on key economic and trade issues and confirmed that AI was among the topics discussed. The BBC said it had contacted the Chinese embassy in the U.S. for comment. In a post on X, Bessent said the talks help lay the groundwork for Trump to advance U.S. economic interests. The two leaders are scheduled to meet Wednesday in Washington.

The context

The New York meetings follow a period of intensifying competition between the two countries over AI and technology. When Xi hosted Trump in Beijing in May, AI collaboration was among the topics the leaders discussed, and those talks produced an agreement for China to buy Boeing jets and U.S. agricultural goods in exchange for lower tariffs. A truce in the tariff dispute between the two economies is set to expire Nov. 10.

Warnings about uncontrolled AI have added urgency to the summit agenda, The Hill reported, with researchers and prominent industry figures recently urging a more cautious approach to development. Trump has largely dismissed those concerns, arguing that existing policies are sufficient and that the U.S. cannot cede its lead to China. "Whoever wins in AI, wins," he said earlier this month, according to the BBC. Beijing, for its part, has resisted a slowdown it views as locking in a U.S. advantage, The Hill reported.

Why it matters

The proposal would create a formal channel between the two largest AI powers for flagging incidents with national security implications, a shift Bessent characterized as moving from opacity toward transparency. Whether it advances could shape how the two governments manage risks from a technology both are racing to dominate. The summit also carries direct trade stakes: the tariff truce expires Nov. 10, and officials discussed a process for identifying possible tariff cuts. Reuters reported the leaders arrive on uneven footing, with China's trade growing while Trump contends with declining approval ratings.

What’s next

Trump and Xi meet Wednesday in Washington, where the AI safety mechanism, tariffs and trade are expected to be on the table. It remains unclear whether China will agree to the notification mechanism; Beijing has not publicly committed to it, and Xinhua's statement confirmed only that AI was discussed. The Nov. 10 expiration of the tariff truce gives both sides a near-term deadline. How the Board of Trade process for tariff cuts takes shape is another open question.