Which concept encompasses the comprehensive Testing, Evaluation, Verification, and Validation activities for AI systems?

• A. Secure by design in AI architecture
• B. Data quality in AI security
• C. Explainability in AI
• D. TEVV in AI system development

Answer: (D)

TEVV, standing for Testing, Evaluation, Verification, and Validation, covers the full set of activities used to assess an AI system from development through deployment. It ensures both conformance to design (verification) and fitness for purpose (validation), while also evaluating performance, safety, and risk in real use. In practice, TEVV encompasses data quality checks, model performance testing, robustness and safety assessments, fairness considerations, integration testing, security-oriented checks, and ongoing post-deployment monitoring. This broad, lifecycle-spanning approach is why it’s the best answer: it goes beyond single concerns like secure-by-design architecture, data quality for security, or explainability, and addresses the complete spectrum of testing, evaluation, verification, and validation activities.