QuestionQ83

VerifiedDomain IV — Understanding how to govern AI deployment and use

Which of the following actions should the company implement to most effectively reduce its risk of reputational damage from use of the AI tool?

Explanation

Pre-deployment testing can identify discriminatory, inaccurate, or unreliable behavior before an AI hiring system affects applicants, while post-deployment testing and monitoring can detect performance drift, unanticipated impacts, and new compliance risks in actual use. Continuous testing therefore directly mitigates the harms most likely to cause reputational damage. NIST’s AI Risk Management Framework calls for test, evaluation, verification, and validation processes and post-deployment monitoring across the AI lifecycle.

Learn more

Community Discussion

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