What makes AI Interpretable to a Human? [Lecture]