The Western AI safety playbook doesn’t transfer to China by jacquesthibs

The author argues that Western AI safety frameworks and funding models do not transfer directly to China due to legal restrictions on foreign nonprofits, different state-backed institutional structures, and geopolitical suspicions. The piece examines how safety work in China is instead driven by state affiliates, universities, and social enterprises, while noting that tightening compute constraints and distillation efforts are reshaping how Chinese labs approach the AI race. Open article →