A comprehensive framework for evaluating the safety and accountability of advanced AI systems, comprising harmonized terminology, a taxonomy for evaluating AI components and systems, and a mapping to the AI system lifecycle and stakeholders.
