Policy Design CRAGE Practice Questions (Page 6)

Questions 26–30

1. 26expert · hard

   A large enterprise has an AI governance policy that includes mandatory bias testing for all AI models. However, the policy is not consistently enforced because different departments interpret the requirements differently. The compliance team wants to improve enforcement without creating excessive bureaucracy. Which approach is most effective?

   • AReplace bias testing with a simpler requirement, such as documenting the model's intended use.
   • BIncrease the frequency of audits without providing additional guidance or tools.
   • CAllow each department to define its own bias testing procedures as long as they report results quarterly.
   • DCentralize bias testing in a dedicated AI ethics team that provides standardized tools and clear pass/fail criteria for all departments.
   Reveal Solution Discussion Select an answer first
2. 27foundation · easy

   Which stakeholder engagement method is most effective for gathering diverse perspectives during AI policy design?

   • AInternal memo circulation
   • BPublic consultations and workshops
   • CExecutive-only review
   • DLegal compliance checklist
   Reveal Solution Discussion Select an answer first
4. 28expert · hard

   A national health service is designing an AI governance policy for a patient risk prediction system. The policy must incorporate input from clinicians, patients, and data scientists, but these groups have conflicting priorities: clinicians want clinical autonomy, patients want transparency, and data scientists want to protect proprietary algorithms. The policy team has limited time and budget. Which stakeholder engagement strategy best reconciles these conflicting interests?

   • APrioritize the clinicians' perspective, since they are the primary users of the system.
   • BConduct separate consultations with each group and then have the policy team make the final decision without sharing the results.
   • CUse an online forum for all stakeholders to discuss the policy, with no facilitation or structured process.
   • DConvene a multi-stakeholder working group with representatives from each group and use a structured decision-making framework to reach consensus.
   Reveal Solution Discussion Select an answer first
5. 29expert · hard

   An insurance company is designing an AI governance policy for a claims processing system. The system has been classified as high-risk under the EU AI Act. The policy must include a risk management system, but the company also needs to control costs. The risk assessment has identified a potential for biased outcomes, but the mitigation would require significant investment. Which policy decision best balances regulatory compliance with cost control?

   • AReclassify the system as low-risk to avoid the compliance burden.
   • BImplement the bias mitigation and document the investment as a necessary cost of regulatory compliance.
   • COutsource the risk management to a third-party vendor to reduce internal costs.
   • DDefer the bias mitigation and document the risk in the policy without taking action.
   Reveal Solution Discussion Select an answer first
6. 30foundation · easy

   In policy design, what is the purpose of integrating a risk assessment framework?

   • ATo automate all AI decision-making
   • BTo guarantee that AI systems are profitable
   • CTo replace human oversight
   • DTo identify, evaluate, and mitigate potential harms from AI systems
   Reveal Solution Discussion Select an answer first