The process of confirming that an AI system's outputs meet clinical standards and are safe for use in healthcare.