US Treasury Secretary Bessent says he will propose to China the establishment of an AI incident notification channel

U.S. Treasury Secretary Scott Bessent said in an interview with Axios that the United States will propose to China a notification process for AI incidents and expects Beijing to agree. He said he had discussed AI safety with Vice Premier He Lifeng before Xi Jinping's state visit in September, and the notification channel was part of that, with risks discussed including out-of-control AI agents and non-state actors using the technology for cyber or biological attacks. Bessent stressed that the mechanism is limited in scope and is not a security treaty, common standards, or an inspection regime. He believes China's change in attitude is related to its open-source model capabilities and mentioned that Moonshot AI's Kimi model had sent sensitive Chinese data to Anthropic. On domestic regulation, he said government review of frontier models remains voluntary, but the government retains the right to intervene when labs ignore safety concerns, with responsibility borne by lab personnel. To executives who recently warned of possibly losing control of systems, he responded that they should slow down and called such statements alarmism lacking solutions. Background: The White House's September 25 and China's Ministry of Foreign Affairs' September 26 outcome documents already announced that both sides agreed to establish bilateral communication channels for related incidents; the proposal wording in the report comes from the interviewee. Official documents do not specify the counterpart agency or notification thresholds, so it cannot be confirmed that the channel is actually operational.