1 article on ai transparency.
Healthcare AI requires explainability. 'The model said so' is not a clinical rationale. Mechanistic interpretability is the research field trying to change that. What it offers practitioners today, where the gap is, and what to do in the meantime.
