We're now building some of the same superiority complex into commercial AIs to effectively refuse any wrongspeak.