US and China discuss AI safety plan ahead of Trump-Xi summit
BBC News ● Covered by 2 sources
US and China talked about an AI incident hotline before the Trump-Xi summit. It’s a rare bit of cooperation in a race both sides say they can’t afford to lose.
Based on reporting by BBC News — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Top US and Chinese officials have been discussing a new way to flag AI incidents that could touch national security, according to Treasury Secretary Scott Bessent. He said the idea came up in New York during what he called successful talks with Chinese Vice Premier He Lifeng.
The timing matters. Donald Trump and Xi Jinping are set to meet in Washington later this week, and the two governments are still locked in a wider contest over trade, chips and AI. Bessent said the goal was to move from opacity to more transparency between what he called the world’s number one and number two AI powers.
This all comes after a stretch of louder warnings about AI’s risks. Researchers have been raising alarms, and some of the industry’s most visible figures have been urging developers to slow down and be more careful. Trump, by contrast, has brushed off the caution and argued the US has enough rules already. He’s also framed the race in blunt terms: whoever wins in AI, wins.
The talks were not just about AI. Bessent said they also covered a process to identify possible tariff cuts on goods known as the Board of Trade. That matters because the current truce in the tariff fight between the two biggest economies is due to expire on 10 November. Chinese state media said the meeting was candid, in-depth and constructive, and that AI was among the topics discussed.
My take — AI-written commentary, not fact-checked reporting
This is the sane part of AI diplomacy: less chest-thumping, more notification mechanisms. The strange bit is that the world’s two biggest AI powers still need a reminder that not every model failure should become a geopolitical surprise. Closed-door rivalry is easy; basic incident-sharing is the grown-up move.
Read more about this at: BBC News
