AI Risk Metrics, Monitoring, and Reporting AAIR Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   What is the primary benefit of using automated tools for AI risk data collection and analysis?

   • AThey reduce the need for data governance practices
   • BThey guarantee that all AI risks are completely mitigated
   • CThey enable real-time or near-real-time monitoring of AI risk indicators
   • DThey eliminate the need for human involvement in risk management
   Reveal Solution Discussion Select an answer first
2. 22foundation · easy

   Which of the following events should trigger a review of AI risk metrics?

   • AAn increase in the number of users without other changes
   • BA routine password change by an administrator
   • CA significant change in the AI system's functionality
   • DA minor update to the user interface
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A government agency's AI risk team must report to regulators on the performance of an AI system used for benefit eligibility. What is the most important element to include in the report?

   • AA detailed explanation of the AI model's architecture and training data
   • BA comparison of the AI system's performance to competitors
   • CEvidence of compliance with regulatory requirements and risk management practices
   • DA list of all employees who have access to the AI system
   Reveal Solution Discussion Select an answer first
5. 24application · medium

   A logistics company uses AI to optimize delivery routes. The risk team wants to automate the collection of data on route deviations and delivery delays. Which approach is most effective?

   • AManually collect data from driver reports at the end of each day
   • BAsk the IT team to export data from the AI system weekly
   • CUse a spreadsheet that drivers update manually
   • DUse APIs to automatically pull data from the AI system and delivery tracking systems into a central monitoring platform
   Reveal Solution Discussion Select an answer first
6. 25expert · hard

   A healthcare AI system monitors patient deterioration. The risk team wants to implement continuous monitoring but is concerned about alert fatigue from false alarms. What is the most effective approach to balance timely detection with reducing alert fatigue?

   • AUse a combination of thresholds and anomaly detection algorithms to reduce false positives while maintaining sensitivity
   • BSet a very high threshold to only alert on severe issues
   • CDisable alerts and rely on manual review of the monitoring dashboard
   • DSet a very low threshold to catch all potential issues
   Reveal Solution Discussion Select an answer first