[CHINA] on AI 中国智能 · THE OTHER RIVER
← The China Pulse
policy · July 24, 2026

China's Vice FM Flies to Washington to Sound Out U.S. on AI-Talks Agenda中国副外长赴华盛顿,摸底中美AI对话议程

中文摘要

Bloomberg reported Friday that Chinese officials are actively trying to clarify what Washington expects from upcoming bilateral AI talks, as the Trump administration maintains escalating pressure on China's AI sector. Vice Foreign Minister Ma Zhaoxu traveled to Washington this week with a remit that includes sounding out U.S. agencies on the scope and likely outcomes of any dialogue.

Beijing is, in part, trying to determine the appropriate seniority and composition of its own delegation before committing to a format. The discussions remain informal — no date or venue for formal talks has been set — and people familiar with the matter requested anonymity to describe the private soundings.

The diplomatic probe comes against a backdrop of contested export controls, the APEC Chengdu statement on open-source AI signed earlier this week, and ongoing U.S. scrutiny of Chinese model deployments. The gap between the two sides' stated positions on AI governance remains wide.

Updated · 1:00 PM EDT · Fri, Jul 24

Bloomberg reported Friday morning that Chinese officials are actively seeking clarity from Washington on what the U.S. expects from planned bilateral AI talks, as the Trump administration simultaneously escalates pressure on Chinese labs over alleged distillation and export-control violations. The new development: analysts now say the two tracks — sanctions and safety dialog — are in direct conflict. Paul Triolo at Albright Stonebridge warned that depending on how many Chinese companies are targeted, retaliation 'has the potential to scuttle both the AI dialog and the September 24 meeting between Presidents Trump and Xi.'

The sanction threat stems from White House OSTP director Michael Kratsios's Wednesday accusation that Moonshot ran a covert platform to distill Anthropic's Fable 5 for Kimi K3, compounded by Treasury Secretary Scott Bessent warning that 'sanctions and Entity List designations will be on the table.' Beijing, per Reuters, is separately weighing whether to restrict overseas users' access to Chinese models as a retaliatory measure.

No formal designations have been announced as of Friday. But the simultaneous escalation on both the sanctions and the dialog tracks — with a presidential summit nominally on the calendar — makes the next several weeks a high-stakes interval for U.S.-China AI relations.

Updated · 3:00 PM EDT · Fri, Jul 24

Bloomberg reported on Friday that Vice Foreign Minister Ma Zhaoxu visited Washington this week with a remit that includes sounding out US agencies on the scope and likely outcomes of a September AI safety dialogue. Chinese officials want clarity on the agenda before committing delegations — specifically, deciding which ministries and at what seniority level.

The visit comes as the Trump administration escalated rhetoric against Chinese AI labs, with Treasury's Bessent warning of sanctions over Kimi K3's alleged Fable 5 distillation. Analysts cited by Bloomberg say the sanction threats now risk undermining the very safety dialogue both sides agreed to pursue, creating a structural tension between the IP enforcement agenda and the stability-focused diplomatic track.

US and Chinese officials are scheduled to hold a formal AI summit in September. The window before that meeting is described as unlikely to see hard enforcement action unless the Treasury uncovers provably egregious violations — but the Moonshot accusation has narrowed that window considerably.