Treasury Secretary Scott Bessent says the US proposed a way to alert China about serious AI incidents, as part of a new AI dialogue. China’s account mentions AI talks but not the proposal, and a Trump–Xi summit follows this week.
New to this? Read it in simple words
- US and Chinese officials met in New York on September 20. Treasury Secretary Scott Bessent says the US proposed alerts for serious AI incidents.
- China’s state news agency Xinhua reported the talks but did not mention the alerts. Bessent says officials will meet again in Shenzhen in about two months.
- Bessent says more openness between the two biggest AI powers is important. A working alert line could reduce confusion during a serious AI failure.
- China has not confirmed the plan. No one has published which events count or who would get an alert.
- AI incident
- An event where an AI system fails or is misused and causes harm.
The US proposed alerts for serious AI incidents
Treasury Secretary Scott Bessent and Trade Representative Jamieson Greer met Chinese Vice Premier He Lifeng in New York on September 20. The talks covered trade and artificial intelligence.
Bessent said the US proposed a notification mechanism for AI incidents that reach a national-security level. It would be part of a new US–China dialogue on AI. He said more transparency between the two biggest AI powers is important.
The next day, Reuters reported comments that Bessent made to CNBC. He said the two sides had formalised a dialogue that includes an incident line. He also said officials will meet again in about two months in Shenzhen.
The US proposed alerts for security-level AI incidents. China has not publicly confirmed the plan.
China’s account is more careful
China’s state news agency Xinhua said the two sides held dialogue on issues related to AI. It called the talks candid, in-depth, and constructive. Its report does not mention the US proposal for incident alerts.
Politico reports that one administration official described the channel as mainly Bessent and He Lifeng, not a formal body of experts. Another person dismissed it with a crude word.
According to Reuters, Bessent wants shared rules on the biggest AI dangers, such as uncontrollable agents and cyber attacks by non-state groups. The current trade truce between the two countries ends on November 10.
A useful line needs clear rules
President Trump and Chinese President Xi Jinping are due to meet in Washington on Thursday. AP and NBC News report that AI is a priority topic for the summit.
A crisis channel needs clear definitions. Both sides must agree which events to report, who receives an alert, and how quickly the other side must answer. None of these details has been published.
A working line could reduce confusion during a serious AI failure or cyber attack. For now, readers should watch for any joint text after the summit and for the Shenzhen meeting.
Sources
Every fact in this story comes from the sources below. Open them to check our work.
- 1Research · September 20, 2026Bessent: US proposes AI incident alert system in talks with China Associated Press via ABC News
- 2Research · September 21, 2026U.S. proposes exchanging AI safety alerts with China, Bessent says NBC News
- 3Research · September 21, 2026US, China to meet again on AI safety in two months in Shenzhen, Bessent says Reuters via Yahoo News
- 4Primary source · September 21, 2026China, U.S. have candid, in-depth, constructive exchanges during consultations on economic, trade issues Xinhua
- 5Research · September 22, 2026‘The scientific, technical term is BS’: Inside the new Bessent-China AI hotline Politico via Yahoo News
We used AP, NBC News, and Reuters for Bessent’s statements and compared them with China’s official Xinhua report. We found no US government readout. We describe the incident line as a proposal because China has not publicly confirmed it and its scope is unclear.