US proposes AI dialogue and incident alerts in talks with China

In short: The United States is pushing for a narrow way to talk with China about serious AI risks, but most signs point to only limited, fragile cooperation.

What's going on

Recent reporting suggests neither the US nor China can force the other to accept its preferred approach on AI. Instead, the most realistic near-term outcome looks like a small, careful set of agreements that reduce risk, without resolving the bigger rivalry.

In the latest exchanges, US Treasury Secretary Scott Bessent said Washington proposed a US China AI dialogue. He also described a “notification mechanism” for AI incidents serious enough to raise national security concerns. Think of it like an emergency phone tree, where each side agrees to quickly flag an AI-related event that could spill into a wider crisis.

Reuters reported the idea was discussed with Chinese Vice Premier He Lifeng ahead of a potential Trump Xi meeting. The Associated Press said the proposal is meant to work like an incident alert system for AI events that could affect national security.

At the same time, several outlets said distrust remains the main barrier. US officials worry that cooperation could help China’s military and economy. Chinese officials argue the US often uses “national security” as a reason to slow China’s tech progress.

What to watch

Analysts expect any progress to stay modest and technical, focused on practical risk reduction rather than broad AI rules. Watch whether the Trump Xi meeting produces a clear agreement on how alerts would work, and whether both sides actually use the channel when problems happen.

Source: NYTimes