The US and China, the world's two AI superpowers, finally opened talks this week aimed at managing emerging safety risks. But whether the Trump-Xi summit can produce meaningful global governance depends on how much the fierce rivals are willing to cooperate.
Treasury Secretary Scott Bessent announced both sides discussed a new AI safety notification mechanism — an open line for either country to alert the other when AI systems pose threats or behave unpredictably. However, the proposal arrives at a time when China may not trust the US to have safety conversations that aren't about limiting Chinese capabilities.
Trump has steadily increased export controls blocking China from advanced AI chips, Congress is considering further restrictions, and the Justice Department recently accused six Chinese AI makers of stealing from US frontier labs. Experts describe the proposed alert channel as superficial, with one source mocking it as 'the scientific, technical term is BS.'
Sam Bresnick from Georgetown's Center for Security and Emerging Technology flagged the absence of technical experts as a red flag. 'It's better than nothing, but I don't have a lot of confidence that it's going to make a major difference,' he said. Alvin Wang Graylin, a Stanford digital economy fellow, warned that the US is too caught up in a 'deeply flawed' narrative about winning an 'AI race' that may benefit no one.
Graylin noted that while the US focuses on frontier AI dominance, Chinese open models are exploding globally and may soon exceed 80% market share. 'US-China capability gaps at the frontiers of AI technology typically last only a few months,' he wrote. 'There is no competitive moat that will secure a permanent advantage.' He urged Trump to narrow export controls to genuine military threats and consider enacting reasonable safety regulations to ease growing public fear.
