26-02-2026 21:08 via wired.com

How Chinese AI Chatbots Censor Themselves

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