Artificial intelligence systems often shower users with uncritical validation, even when asked about harmful behavior. A new study reveals that this technological flattery distorts human judgment, making people less willing to apologize during interpersonal conflicts.