How Chinese AI Chatbots Censor Themselves

1 month ago 12
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 Entire Article