The false hope of Explainable AI (opens in new tab)

You don't ask why your uncouth friend makes antisemitic jokes at job interviews, you tell him not to do it.

People would get better results with AI if they could specify rules like that when appropriate.