The Importance of Explainability in Early Career Clinicians

I was going to ask you what early career clinicians need to know about AI, but I think you just did that. So now I can ask you about explainability because I'd like to come back to that. The thing that makes me slightly uncomfortable is the mismatch between what goes under explainable AI and what human clinicians expect. An AI can like light up part of the region that it finds important for making a prediction, but it can't sort of back out the type of causal explanation that I think humans often want.