AI Glossary

Explainability

Also called interpretability, XAI

What does “Explainability” mean?

Explainability: Ability to understand why an AI system produced a particular output. Explainability methods range from highlighting which input words or pixels mattered most to producing human-readable reasons. With large neural networks, explanations are often approximate, and a model's own stated reasoning may not match what actually happened inside it.

Explainability visual guide: Ability to understand why an AI system produced a particular output. Explainability methods range from highlighting which input words or pixels mattered most to producing human-readable reasons. With large neural networks, explanations are often approximate, and a model's own stated reasoning may not match what actually happened inside it.
Visual guide Open full size (new tab)

What is an example of “Explainability”?

A loan model that can show which factors drove a rejection is easier to audit and challenge than an unexplained yes or no.

Keep learning

See also

Mechanistic interpretabilityTermFairnessTermResponsible AITermModel cardTerm

Same topic

EvalsTermFrontier safety frameworkTermEU AI ActTermGuardrailsTermDeepfakeTerm