Explainability
The degree to which a model's decision can be understood and explained in human terms, a core requirement for AI governance and consumer recourse.
Explainability, sometimes called explainable AI (XAI), is the ability to describe how a model arrived at a particular decision in terms that humans can understand. It is different from accuracy. A model can be accurate and still opaque.
Explainability is central to AI governance because consumers, regulators, and internal reviewers must be able to understand why a decision was made. Without it, human reviewers cannot meaningfully override AI outputs, consumers cannot appeal adverse decisions, and examiners cannot assess fairness.
Improving explainability usually involves simpler model architectures, feature importance analysis, documented decision logic, and good data lineage. See our glossary entries on data lineage, human-in-the-loop, and the AI Systems Program.