US and China Discuss AI Security Alerts as Reports Split on a Deal
Scott Bessent says Washington and Beijing are discussing mutual alerts for AI national-security incidents, but outlets disagree on whether any mechanism is already agreed.
US and Chinese officials have discussed a mutual notification mechanism for AI incidents with national-security implications, which WIRED calls the 'US China AI Dialogue,' according to Treasury Secretary Scott Bessent. WIRED describes newly opened talks that resumed after President Trump's May visit to Beijing and ahead of Xi Jinping's trip to Washington, while The Decoder says the two sides agreed to establish an official AI dialogue and notification system ahead of a Trump-Xi summit. SecurityWeek is mixed: its headline and summary say the United States proposed an alert mechanism, but its points also say the sides agreed to an AI risk notification mechanism. WIRED and The Decoder say US export controls on advanced AI chips were not discussed and stay off the formal agenda, which sits in tension with SecurityWeek's report that the talks cover cross-border AI transparency, trade tariffs, and possible sensitive trade items. Bessent has framed the effort as a move from opacity toward transparency between the top two AI powers. Only WIRED adds that Sam Altman, Dario Amodei, and Elon Musk called for a development slowdown, which the Trump administration and Nvidia's Jensen Huang rejected.
- Treasury Secretary Scott Bessent is tied to a proposed US-China mechanism, called the 'US China AI Dialogue' by WIRED, for mutual notice of AI incidents with national-security implications.
- WIRED says talks resumed after President Trump's May visit to Beijing and before Xi Jinping's trip to Washington; The Decoder places the effort ahead of a Trump-Xi summit.
- Status is disputed: WIRED describes newly opened negotiations; The Decoder says the sides agreed to establish an official dialogue and notification system; SecurityWeek's headline says the US proposed the system while its points also say a…
- WIRED and The Decoder say US export controls on advanced AI chips to China were not discussed and remain off the formal agenda.
- SecurityWeek alone says the talks also cover cross-border AI transparency and trade tariffs, with possible inclusion of sensitive trade items.
- Only WIRED reports that Sam Altman, Dario Amodei, and Elon Musk urged an AI development slowdown, which the Trump administration and Nvidia's Jensen Huang rejected.
- WIRED and The Decoder published on 2026-09-21; SecurityWeek published on 2026-09-22. WIRED refers to weekend meetings.
Coverage timelineoldest first · each row is one article
- · 5d agoUS and China Discuss Alerting Each Other to AI National Security Threats
WIRED · Security· 65
US and China begin talks on a mechanism to mutually notify each other of AI incidents that could threaten national security.
- · 5d agoUS and China agree on AI dialogue with security mechanism ahead of Trump-Xi summit
The Decoder· 75
US and China establish joint AI dialogue with security notification mechanism ahead of Trump-Xi summit.
- · 5d ago