Model Interpretability vs. Explainability

Interpretability means a human can grasp a model's overall logic, like reading a simple decision tree's rules. Explainability is stronger: it's about understanding *why* the model made a specific prediction, with context. This is critical for debugging, building trust, or justifying high-stakes decisions in finance and medicine. The footgun: using the terms interchangeably. A model can be interpretable (simple) but still fail to provide a useful explanation for a specific, critical prediction.