Sustaining AI Transformation and Continuous Improvement CAIPM Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A city government deployed an AI system to allocate resources for public services. The system has been running for a year, but the city council is concerned about fairness and accountability. The system's decisions are not always well-received by the public. The program manager wants to sustain the initiative. What is the most important action to take?

   • ALimit public access to information about the system to avoid criticism
   • BFocus on improving the system's technical performance to address public concerns
   • CIncrease the system's automation to reduce human intervention
   • DEstablish a governance committee that includes community representatives to oversee the system
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   An e-commerce company's recommendation engine is underperforming in terms of click-through rate (CTR). The team has a backlog of potential improvements, including new features, algorithm changes, and UI tweaks. They want to adopt a continuous improvement approach to test changes quickly and learn from user feedback. Which practice best aligns with this goal?

   • AImplement all backlog items at once in the next release to maximize impact.
   • BRun A/B tests for each proposed change and use the results to decide which improvements to keep.
   • CChoose the improvement that the most senior engineer recommends and deploy it immediately.
   • DWait for the next quarterly review to decide on any changes.
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A telecom company has an AI model that predicts customer churn. The model is used by the retention team, but the team rarely acts on the model's recommendations because they do not trust the output. The model's accuracy is high, but the team prefers their own judgment. What should the program manager do to improve the use of the model?

   • AProvide training to the retention team on how the model works and how to interpret its outputs
   • BReplace the retention team with an automated system that acts on the model's recommendations
   • CIncrease the model's accuracy threshold to make recommendations more reliable
   • DDiscontinue the model because the team does not use it
   Reveal Solution Discussion Select an answer first
5. 14expert · hard

   A manufacturing firm's AI quality-inspection system is technically successful, but operators in one plant have stopped using it because they distrust the system's recommendations. The plant manager wants to restore trust and ensure sustained use. The system's false-positive rate is higher than in the pilot plant. What should the program manager do first?

   • AReplace the system with a different AI vendor's product.
   • BProvide additional training to operators on how the system works.
   • CInvestigate the root cause of the higher false-positive rate and adjust the system's threshold or retrain if needed.
   • DMandate that operators use the system and report non-compliance to senior management.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A financial services firm has a credit-risk model that must be retrained quarterly due to changing economic conditions. The compliance team requires a complete record of every model version, its training data, and its approval status. The data science team currently saves model files with timestamps but lacks a formal process. What should the team implement to meet the compliance requirement?

   • AA quarterly meeting to review and approve model changes manually.
   • BA model registry that tracks versions, metadata, approval status, and audit logs.
   • CA shared network folder with read-only permissions for all model files.
   • DA backup system that stores nightly snapshots of the model files.
   Reveal Solution Discussion Select an answer first