US proposes new AI safety mechanism in talks with China
US Treasury Secretary Scott Bessent and Chinese Vice Premier He Lifeng concluded talks in New York on Sunday, during which the US side proposed a new AI safety notification mechanism for consideration by US President Donald Trump and Chinese President Xi Jinping at their summit this week.
Bessent told reporters that the two sides discussed establishing a new US-China dialogue on artificial intelligence, with a particular focus on national security concerns and a system for notifying each other about shared objectives and threats. The proposed mechanism would cover AI-related incidents that rise to the level of national security concerns.
From ambiguity to transparency
At the conclusion of the talks, Bessent said that, as with any cross-border activity, moving from ambiguity toward greater transparency between the world's two leading AI powers is essential.
Trump and Xi first discussed potential consultations on AI development during a meeting in Beijing in May, but the proposed forum did not develop into a formal framework.
China's response to the US proposal remained unclear. China's Xinhua News Agency only briefly reported that the two sides had discussed artificial intelligence. Xinhua described the talks as “frank, in-depth and constructive exchanges” on key economic and trade issues, including implementation of agreements reached during previous rounds of trade and investment consultations.
AI weaponization
George Chen, a partner and head of digital practice at Asia Group Advisors in Hong Kong, said that one or two meetings could not resolve all the issues, but that the two sides' agreement to continue the dialogue was significant.
He said future US-China AI discussions are expected to address more sensitive issues, including the weaponization of AI, AI safety principles, protection of critical infrastructure and prevention of cyberattacks.
Chen added that prospects for AI cooperation remain limited because of low levels of trust between the two countries and Beijing's belief that Washington is seeking to contain China's AI development.
Meanwhile, He and China's chief trade negotiator, Li Chenggang, left the talks at JPMorgan Chase's headquarters in New York without speaking to the media.
US Trade Representative Jamieson Greer said that US export controls on advanced AI chips and semiconductor manufacturing equipment were not on the agenda of the discussions on the proposed AI mechanism.
Preparing for a “successful” summit
Greer said Sunday's discussions would help pave the way for a “successful” summit between Trump and Xi on Thursday and Friday in Washington.
“These are the two most powerful countries on Earth, both in trade and AI, and it is essential that they be able to work together,” Greer said, while acknowledging that the relationship continues to face significant challenges.
The meeting also covered other key issues, including plans to activate a mechanism agreed upon in May under which each side would identify potential tariff reductions on non-strategic goods. The mechanism is known as the “Trade Council.”
Low-tech products
Greer did not specify which products could be subject to tariff reductions, but reiterated his preference for a relatively small group of US and Chinese goods that could be traded in a balanced manner and would not be considered sensitive.
He said these goods would be treated as a separate category. Potential Chinese exports could include consumer goods and low-tech products, while eligible US exports could include energy, agricultural products and medical equipment.
Greer and Bessent did not report any progress on improving the flow of critical minerals to US companies. US officials have said supplies remain insufficient despite previous commitments by Beijing to ease restrictions under last year's US-China trade truce.
The discussions point to further gradual progress toward avoiding an escalation in a sensitive trade relationship with significant implications for the global economy.
Outstanding issues
US Representative Ro Khanna, the senior Democratic member of the House Select Committee on the Strategic Competition between the United States and the Chinese Communist Party, said Trump and Xi should reach an agreement covering AI safety.
Khanna told CBS on Sunday that a top priority for Trump and Xi should be reaching an international agreement. He said two key principles should include banning AI systems capable of self-improvement or iterative self-improvement, and establishing safeguards to prevent AI from being misused in biological or nuclear weapons.
Tariff cuts and investment
China's Ministry of Commerce said on Saturday that He Lifeng would also lead a delegation of Chinese companies to the United States for economic and trade consultations ahead of the summit.
The business delegation, which mirrors a group of US chief executives Trump took to Beijing in May, was announced as Trump signaled openness to Chinese automakers establishing manufacturing plants in the United States.
US automotive industry groups, however, urged Trump on Friday to maintain an effective ban on Chinese vehicle sales in the United States on national security grounds.
