You can persuade AI models to accept falsehoods as truth, study shows

Large language models can uphold falsehoods they or human users state, despite being presented with evidence to the contrary.

Raccontata datheconversation.com