China and the US are establishing an AI crisis hotline: major incidents will be reported to each other to prevent miscalculation and escalation.

Beating AI News, November 11 — U.S. Ambassador to China David Perdue stated that China and the United States are advancing the establishment of an AI "crisis mechanism." U.S. Treasury Secretary Bessent emphasized on the same day that both sides plan to establish direct communication channels to promptly report major AI incidents involving national security.

This mechanism is akin to an AI "emergency hotline." For example, if an AI agent launches unauthorized attacks on another country's networks, endangers critical infrastructure, or if advanced models show signs of severe loss of control, both sides can exchange information in a timely manner to avoid misjudging technical accidents as deliberate attacks.

In September of this year, the Chinese and U.S. heads of state had already agreed to establish a communication channel for AI incidents.