The United States has proposed establishing a mutual AI safety alert system with China, with Treasury Secretary Scott Bessent announcing the initiative ahead of the UN General Assembly gathering where a Trump-Xi meeting is expected. The proposal would create a channel for the two governments to notify each other of significant artificial intelligence developments or safety concerns.

Bessent outlined the notification system concept publicly, framing it as a mechanism to reduce the risk of miscalculation between the world's two leading AI powers. The proposal comes as both countries continue to invest heavily in artificial intelligence research and military applications, raising concerns about the potential for unintended escalation.

The initiative is being floated as a confidence-building measure rather than a formal regulatory agreement, similar in concept to the nuclear hotlines established during the Cold War. Officials have emphasized that the proposal does not involve sharing proprietary technology or intelligence, but rather creating a structured communication pathway for AI-related safety incidents.

The proposal has generated significant debate about whether engagement with Beijing on AI governance is strategically sound. Critics argue that any cooperative framework with China on AI could inadvertently legitimize Chinese AI practices or expose American vulnerabilities, while proponents contend that the absence of communication channels poses its own serious risks as AI systems grow more capable and integrated into critical infrastructure and defense systems.