Explainability (XAI)

intermediate

The ability to understand and explain how an AI system reaches its decisions. Required by the EU AI Act for high-risk systems: users must be able to understand why an AI made a specific recommendation or decision. Techniques include feature importance, attention visualization, and counterfactual explanations.