How Chinese AI Chatbots Censor Themselves

Wired Business
Generative AI

Researchers from Stanford and Princeton found that Chinese AI models are likely than their Western counterparts to dodge political questions or deliver inaccurate answers.