Trump's China Trip Unlikely to Yield AI Progress
Original: Even His Own Officials Don’t Think Trump’s China Trip Will Lead to Progress on AI
Why This Matters
No working AI safety channel between the world's two largest AI powers is a serious structural gap.
Ahead of Trump's meeting with Xi Jinping, U.S. officials privately say little progress is expected on AI safety or espionage concerns, with the Nvidia scandal further straining already difficult negotiations between the two powers.
Administration officials told Wired's Inner Loop they see little room for progress on AI issues during Trump's summit with Xi Jinping on Thursday. Trump himself left open whether he would raise Chinese interference in AI or the Nvidia scandal, as officials say the administration is trying not to antagonize Beijing.
The fallout from former Nvidia CEO Jensen Huang's alleged affair with Alice Huang—an executive at Megaspeed and suspected Chinese intelligence asset—has made AI negotiations especially tense. A planned AI incident hotline between Treasury Secretary Scott Bessent and Chinese vice premier He Lifeng, announced in September, remains non-operational. Secretary of State Marco Rubio acknowledged last week that AI safety talks had occurred at a "technical level" but called the effort "still a work in progress."
The two sides also hold fundamentally different views on AI risk. U.S. officials worry about autonomous, misaligned AI models operating beyond human oversight. China is reportedly more concerned about advanced open-weight models being used by dissidents to undermine CCP authority—and is suspected of using frontier AI for domestic and foreign intelligence gathering, making voluntary incident reporting unlikely. Meanwhile, the Trump administration this week announced a new rule limiting China's ability to remotely access U.S.-made AI chips in foreign data centers, a direct response to Megaspeed's alleged chip access.