U.S. and China to Hold First AI Safety Talks in Beijing in Mid-September; Bessent to Lead, China Previously Set Definition Prerequisite

U.S. Treasury Secretary Bessent will lead a delegation to Beijing in mid-September 2026 for the first dedicated bilateral AI safety dialogue with China during Trump's second term.

Facts

According to Bloomberg, Reuters, CNBC and other independent sources, the dialogue is scheduled for mid-September, with U.S. Treasury Secretary Bessent leading the U.S. side. Topics include risks of autonomous AI cyberattacks and agreements on data sharing among frontier labs. China had previously set a precondition: the two sides must first reach a consensus on the definition of “AI safety,” otherwise negotiations would not proceed. The context is the May Trump-Xi summit that restarted the AI dialogue mechanism, with the next milestone being another Trump-Xi meeting in Washington on September 24. This is the first dedicated AI safety consultation since the May 2024 Geneva meeting.

Mechanism Breakdown

This dialogue is positioned as the first bilateral mechanism specifically focused on AI safety between the two countries during Trump's second term. The U.S. side will be led in talks by Treasury Secretary Bessent, indicating cross-cutting considerations of finance and technology security. China's precondition on the definition issue reflects that the two sides still differ on core concepts. The meeting directly follows the outcomes of the May summit and points toward higher-level contact on September 24, forming a continuous chain of policy advancement.

Industry Impact

The launch of negotiations on AI safety boundaries will directly affect global frontier labs' data-sharing rules and cross-border cooperation pathways. The risk of autonomous AI cyberattacks has become a core topic, meaning constraints on model behavior may rise from technical standards to inter-state agreements. Companies need to watch for potential adjustments in future data flows and security reviews, especially compliance costs involving cross-border model training and deployment.

Strategic Assessment

[Analysis] Based on available facts, the sensitive timing of this dialogue and China's preconditions show that the two sides remain at the stage of conceptual alignment on AI safety rather than substantive agreement negotiations. The subsequent high-level contact on September 24 may determine whether the mechanism can move from dialogue to an executable framework, but current points of divergence will still constrain the pace of progress.