US-China talks on AI safety offer no chance of progress, administration officials say
Officials in the Trump administration see only a slim chance of reaching an agreement with China on AI safety. The planned AI incident hotline still isn't functioning, the two sides have different concerns, and mutual distrust over suspected espionage is deepening.
Ahead of President Trump's meeting with Xi Jinping, officials of the US administration privately told Wired that they see no room for progress in talks with China on AI safety and espionage. The AI incident hotline between US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng, agreed upon back in September, remains only a concept and is not functional. Secretary of State Marco Rubio confirmed that talks are taking place only at a "technical level" and are "still a work in progress."
According to officials, the two countries disagree on what constitutes the biggest risk from AI. The Trump administration and American technology companies are concerned about autonomous and misaligned AI models operating beyond human oversight. The Chinese government, on the other hand, fears that advanced open-weight models could be used by critics and dissidents to undermine the Communist Party of China. At the same time, officials consider the Chinese government to be an actor that wants to use frontier AI models for intelligence activities at home and abroad, which reduces its willingness to report incidents where its models slip out of control.
Talks are further complicated by the case of former Nvidia CEO Jensen Huang and his relationship with Alice Huang, a manager at the company Megaspeed, whom officials suspect of ties to Chinese intelligence services. China has rejected these accusations as "unfounded" and a "witch hunt." Despite President Trump's efforts toward friendlier economic relations with China, the administration this week announced a new rule restricting China's remote access to American AI chips in foreign data centers — a direct response to Megaspeed's role in the Nvidia affair, which could also affect similar companies with Chinese owners.
Why it matters
A non-functioning communication channel between the US and China increases the risk that an AI model incident (loss of control, misuse, espionage) will not be reported or addressed in time, since both sides have different motivations for disclosing problems. New regulation restricting remote access to American AI chips in foreign data centers has a direct impact on companies with ties to China that operate infrastructure using Nvidia chips.
Relevant practical impact
What this means
For a business
Companies operating foreign data centers with US AI chips, especially those with ties to Chinese capital, face new regulation restricting remote access to these chips, which may require reviewing supply chain and ownership links.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.