The US and China increasingly recognise similar frontier AI risks, but differ sharply on how and when regulators should ...