China's spy chief warns AI hacking risks threaten national security

China's spy agency chief warned this week that advanced AI models' hacking capabilities could jeopardise national security, shifting Beijing's focus from immediate risks like deepfakes to broader threats. The warning follows a California team's demonstration of an AI-built cyberattack "worm" that could compromise millions of WeChat accounts. Chinese policymakers now see AI safety issues as more similar to US concerns than previously acknowledged, though such debates rarely go public in China. The cybersecurity impact of autonomous AI agents ranks among the government's top worries, prompting pressure on Presidents Xi Jinping and Donald Trump to address the threats at Thursday's talks. Beijing treats AI as a manageable general-purpose technology, like electricity, and already regulates uses such as unauthorised AI avatars. However, cooperation on shared safety evaluations may be undermined by competition, with Trump vowing to maintain US leadership and China fearing "security" could restrict its technological development.