explainabilityThe ability to understand and assess how an AI model reaches its outputs, supporting regulatory oversight.8 September 2026