US officials see little chance of AI progress at Trump-Xi summit

Officials told WIRED's Inner Loop newsletter that a planned AI incident hotline is still only a concept and that mistrust over chips and espionage is clouding talks with China.

· 2 min read

people in conference
Photo: Evangeline Shaw / Unsplash

Officials in the Trump administration expect little progress on AI safety or AI-related espionage when President Donald Trump meets Xi Jinping on Thursday, according to WIRED. The outlet's Inner Loop newsletter reported that the officials spoke privately and see limited room for breakthroughs.

WIRED said it was unclear whether Trump would raise Chinese interference in AI, or a scandal involving Nvidia, with Xi. Officials described the president's approach as an effort not to antagonize Beijing during the summit.

The Nvidia matter involves Megaspeed, a firm whose executive is suspected of being a Chinese intelligence asset. The report says the episode has made AI negotiations between the two largest economies more fraught. China has called the allegations groundless and a smear.

In September, the administration moved to create an AI incident hotline between Treasury Secretary Scott Bessent and Chinese vice premier He Lifeng. According to the report, the channel is still at the concept stage and does not work yet.

Secretary of State Marco Rubio told reporters last week that the two governments had held AI safety talks at a technical level since the hotline was announced. He conceded the effort was still a work in progress. Officials have also discussed a framework setting out when each side must alert the other, with enforcement so the US is not the only one sending warnings.

Officials told WIRED that the two countries disagree about the main risks. US officials and tech leaders worry most about AI models becoming autonomous and misaligned, operating beyond human oversight. Chinese officials are widely understood to be more concerned that dissidents and critics could use advanced open-weight models to undermine the Chinese Communist Party.

Incentives differ as well. The report says Beijing appears interested in using frontier AI models to gather intelligence at home and abroad, and would be unlikely to disclose cases where its own models went out of control. Mutual suspicion of espionage has deepened the distrust, and China had already accused the US of trying to hold back its AI development before the Nvidia episode.

While Trump has pursued friendlier economic ties with China, his administration this week said it would go ahead with a rule restricting China's ability to remotely access US-made AI chips in data centers abroad. WIRED described it as a direct intervention by Trump in response to Megaspeed's alleged role and to criticism that a firm tied to China had obtained advanced AI chips. WIRED has previously reported that the rule could also keep companies with Chinese corporate parents from doing business with Nvidia.

Related