Can One Chatbot Catch Another’s Lies?

A new approach uses language models to interrogate other language models and sniff out lies