A Penn State study shows that rude prompts yield sharper AI answers, challenging assumptions that politeness improves model accuracy.
