QuestionQ62

VerifiedDomain III — Understanding how to govern AI development

All of the following testing types can help assess the performance of a responsible AI system EXCEPT:

Explanation

Adversarial robustness, statistical sampling, and decision analysis can test a system’s behavior, accuracy, resilience, or decision quality. Risk probability and severity instead characterize and prioritize potential harms according to likelihood and impact; they support risk management rather than directly testing operational performance. Microsoft’s responsible-AI guidance likewise treats severity and probability as factors for evaluating and prioritizing harms.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!