U.S. to propose emergency AI notification system with China

The statement was made in an interview on "The Axios Show." Axios published the report on October 3, 2026, under the headline "U.S. to propose emergency AI notification system with China." Bessent said he thinks Beijing would agree to the arrangement.

Bessent said the talks covered risks including uncontrolled AI agents and non-state actors using the technology for cyber or biological threats. The excerpt describes the notification process as something Bessent plans to propose, rather than as an agreement that has been reached. It does not name the other participants in the talks, state when they took place, or specify what form the notification process would take. It also does not state whether a formal proposal has been delivered to Beijing.

According to the excerpt, the proposed mechanism concerns notification between the two governments if something goes wrong with AI. The risks Bessent listed — uncontrolled AI agents and non-state actors using the technology for cyber or biological threats — involve parties other than the two governments involved in the talks. The report quotes Bessent's expectation that Beijing would agree to the process.