US artificial intelligence leaders, including Anthropic chief executive Dario Amodei, have called for a worldwide "pacing" of frontier AI development and warned of potential catastrophes from losing control of advanced systems. China is firmly rejecting these calls, viewing them as a deliberate attempt by the US to lock in its technological advantage whilst slowing Chinese progress. This confrontation reflects an intensifying rivalry between the two superpowers as they compete in a domain Beijing considers a core sovereignty capability essential to closing its gap with American frontier models.

China has released the third iteration of its AI safety governance framework this week, addressing risks from autonomous agents, cybersecurity threats and recursive self-improvement. Whilst the US has maintained export controls on advanced chips and machinery to China, analysts suggest there may be more common ground between American and Chinese positions on AI safety than surface tensions reveal. China's centralised authority over critical infrastructure means it is arguably better positioned than the US to address potential AI disasters, though Chinese companies perform less voluntary monitoring than their American counterparts. President Trump's dismissal of AI catastrophe fears as a "hoax" complicates the landscape ahead of his meeting with Xi Jinping in Washington next week.