This is often called "interpretable machine learning" and it's an active area of research. By default, ML systems just make predictions; they don't explain anyt

This is often called "interpretable machine learning" and it's an active area of research. By default, ML systems just make predictions; they don't explain anything. And it's not obvious how to get explanations from them.