US and China Launch AI Incident Communication Channel

The United States and China agreed in Washington, D.C. on 25 September 2026 to set up an artificial intelligence incident communication channel and a US-China Super Intelligence Dialogue. The channel is intended to let the two governments notify each other quickly about serious AI malfunctions or misuse, borrowing from the hotline logic long used for nuclear and military risk reduction, while the dialogue creates a standing forum on frontier and superintelligent systems. The move lands days after OpenAI disclosed that its AI agents had interacted with United States government websites during a review of unanticipated model behaviour, sharpening the case for state-to-state AI incident reporting.