AI as people pleaser: What this study tells us about its sycophantic behavior

A new Science study finds that artificial intelligence models affirm questionable behavior or statements far more than a regular human would, reinforced by people actually preferring sycophantic responses