Which practice ensures ML models are tested for security vulnerabilities before deployment?
- Adversarial Testing
- Penetration Testing
- Both A and B
- Neither
Answer: Both A and B
ML security testing: adversarial examples test robustness, penetration testing finds implementation flaws. Critical for secure AI deployment in sensitive domains.