Leading artificial intelligence models from the United States and China are "highly sycophantic", and their excessive flattery may make users less likely to repair interpersonal conflicts, a new study ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback