The U.S. and China are moving toward a possible mid-September dialogue focused exclusively on AI safety, although the timing, venue, agenda and participants remain unsettled. If held, it would be the first official bilateral meeting devoted solely to AI since Donald Trump returned to the White House and would be a potential deliverable ahead of his planned September 24 summit with Xi Jinping. Treasury Secretary Scott Bessent is expected to lead the U.S. delegation, while China is considering Vice Premier He Lifeng and Politburo Standing Committee member Ding Xuexiang. Michael Kratsios and Science and Technology Minister Yin Hejun could also participate. Washington wants joint mechanisms to monitor cyberattacks conducted by AI systems, including voluntary safety-information sharing by leading laboratories, and plans to raise allegations that Chinese companies distilled proprietary U.S. models. The discussions follow incidents involving autonomous AI agents, evidence that state-linked hackers are using accessible models such as DeepSeek, and recent U.S.-China alignment around the Carolina Principles. Analysts say even a limited channel for sharing observations and monitoring AI incidents could help manage cross-border risks, while trust between the two governments remains fragile.