U.S. and China Discuss AI Safety Alert Mechanism

The United States and China have discussed establishing a new artificial intelligence dialogue that could include a notification mechanism for AI incidents reaching the level of national security concerns. Treasury Secretary Scott Bessent disclosed the discussions after meeting Chinese Vice Premier and top trade negotiator He Lifeng in New York on September 20, ahead of a planned meeting between President Donald Trump and Chinese President Xi Jinping.

Bessent and He met at JPMorgan Chase headquarters in New York as the two countries held broader discussions on trade and economic issues. Bessent described the talks as “very successful” and said Washington and Beijing were working toward an “U.S.-China AI dialogue” that would include a system for notifying each other about significant AI-related incidents.

According to Bessent, the proposed mechanism would focus on incidents involving artificial intelligence that rise to a national security level. He said greater transparency between the world’s two leading AI powers was important, particularly for activity that crosses borders. Bessent later indicated that officials would likely meet again in about two months to continue the discussions.

China’s state-run Xinhua news agency confirmed that the two sides held discussions involving AI, while describing the broader economic and trade consultations as “candid, in-depth and constructive.” Xinhua said the delegations also discussed important economic and trade matters of mutual concern and the implementation of understandings reached during earlier negotiations.

The AI discussions come as technology executives and researchers debate whether the development of increasingly advanced AI systems should be slowed to address potential safety risks. Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have been among the industry figures involved in recent warnings about the risks associated with rapidly advancing AI.

China has pushed back against calls for a slowdown. On September 14, Chinese Foreign Ministry spokesperson Guo Jiakun said all parties should work together on AI and warned that “fearmongering, confrontation and vicious competition” would disrupt global AI governance. His comments followed concerns raised by Amodei about the implications of China gaining an AI lead.

Trump has meanwhile emphasized continued AI development and competition with China. On September 19, he announced plans to establish an “AI Force” and appoint a new AI czar, while saying his administration would not “hinder or stifle” the growth of the industry. Trump said the existing criminal and civil justice systems could be used to address harmful applications of AI.

The proposed U.S.-China AI notification system therefore comes as Washington and Beijing pursue dialogue on AI risks while continuing to compete over the technology’s development. The issue is expected to form part of the broader agenda surrounding the upcoming Trump-Xi meeting in Washington.