Policy Design CRAGE Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   What is the primary purpose of continuous monitoring in AI policy implementation?

   • ATo detect deviations from policy and performance issues over time
   • BTo guarantee that AI systems never make errors
   • CTo replace the need for initial risk assessments
   • DTo eliminate the need for stakeholder engagement
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   Which ethical consideration is most directly relevant when designing AI policies that affect vulnerable populations?

   • ANetwork bandwidth optimization
   • BUser interface aesthetics
   • CIntellectual property protection
   • DHuman rights and societal impact
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   An e-commerce company is updating its AI governance policy to include a risk assessment framework for a recommendation system. The system has been flagged for potentially amplifying harmful content. The policy team must decide how to integrate risk assessment to address this issue. Which approach is most appropriate?

   • AInclude a specific risk assessment step for content amplification, with a threshold that triggers mandatory mitigation.
   • BConduct a general risk assessment without focusing on content amplification, since the system is not high-risk.
   • CRequire the recommendation system to be reviewed by the legal team only, without technical input.
   • DRemove the recommendation system from the policy scope to avoid overcomplicating the governance framework.
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A manufacturing company uses AI for predictive maintenance. The policy team is integrating risk assessment into the AI governance policy. They identify that a false negative (failing to predict a breakdown) could cause safety incidents, while false positives cause unnecessary downtime. Which risk mitigation strategy should the policy prioritize?

   • AImplement a policy that requires the AI system to achieve 100% accuracy before deployment, otherwise it will not be used.
   • BSet a higher threshold for issuing maintenance alerts to reduce false positives, accepting more false negatives.
   • CDesign the policy to favor fewer false negatives by lowering the alert threshold, and include a protocol for investigating missed predictions.
   • DRequire a human review of all AI-generated maintenance alerts before any action is taken.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A manufacturing company has implemented an AI governance policy that requires human oversight for all AI-driven quality control decisions. The policy is not being followed consistently on the factory floor. Which implementation strategy would most effectively improve compliance?

   • APost the policy on the company intranet and assume employees will read it.
   • BDelegate all oversight responsibilities to a single quality manager without additional training.
   • CRevise the policy to remove the human oversight requirement, since it is difficult to follow.
   • DProvide role-specific training to factory floor supervisors on when and how human oversight is required.
   Reveal Solution Discussion Select an answer first