Washington Seeks New System for Major AI Incidents

The United States has proposed a new AI incident notification mechanism during weekend talks with Chinese officials in New York, Treasury Secretary Scott Bessent said Sunday.

The proposal is aimed at improving communication when artificial intelligence incidents could pose national security risks, including potentially serious problems involving advanced AI systems. The discussions come ahead of an expected meeting between President Donald Trump and Chinese President Xi Jinping.

U.S. and China Discuss AI Risks

Bessent said greater transparency between the world's two leading AI powers could help establish a shared understanding of emerging risks.

The discussions involved Bessent and Chinese Vice Premier He Lifeng, with both sides also addressing broader economic and trade issues. Chinese state media described the AI discussions as candid and constructive.

AI Security Moves Into Global Focus

The proposal comes as governments increasingly examine the security implications of increasingly capable artificial intelligence systems.

A formal incident-notification system could provide a channel for the United States and China to communicate about major AI-related events and potential national-security threats. The details of how such a mechanism would operate have not yet been finalized.

The discussions also included trade, with U.S. officials saying the two countries were considering further talks involving products such as consumer goods, agricultural products, energy and medical devices.

The development adds a new dimension to U.S.-China technology and national-security discussions as both countries prepare for further high-level talks.