Loading…
Glossary term
The degree to which people can understand why an AI system produced a particular output.
Explainability is about giving people an understandable reason for a model's output: which inputs mattered, or which rule or example drove a decision. Tools include feature-importance scores and highlighting the part of an image a model used.
An explanation is useful only if it is faithful to how the model actually worked.
People affected by automated decisions, in lending or hiring for instance, often have a right or need to understand them.