AI Implementation, Maintenance, and Decommissioning AAIR Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   Which of the following is an example of an ongoing maintenance activity for an AI system?

   • ARemoving the model from production permanently
   • BDeleting the model's audit logs after one week
   • CChanging the model's architecture every day
   • DPeriodically retraining the model on new data
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which activity is part of a post-decommissioning review to ensure regulatory compliance?

   • ARe-enabling the system's user accounts
   • BRestoring the system from backup
   • CIncreasing the system's logging level
   • DVerifying that all required data retention and deletion obligations were met
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Which activity is part of verifying that an AI model meets compliance requirements during implementation?

   • AReviewing the model's decisions for alignment with applicable regulations and policies
   • BOptimizing the model's hyperparameters for faster training
   • CSelecting a new cloud provider for hosting
   • DIncreasing the number of training epochs
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   What is a key ongoing activity for maintaining an AI system's performance and reliability?

   • AMonitoring data drift and model performance metrics
   • BRemoving all human oversight to reduce operational burden
   • CFreezing the model to prevent any changes
   • DDisabling logging to reduce storage costs
   Reveal Solution Discussion Select an answer first
6. 5expert · hard

   A streaming service has deployed an AI system that recommends content. The system is monitored using a dashboard that tracks recommendation click-through rate (CTR) and user engagement. After a major sports event, the CTR drops significantly. The data science team suspects the model is not adapting to the temporary change in user behavior. The company wants to maintain service quality without overreacting to short-term fluctuations. What should they do?

   • ARoll back to the previous model version to restore the CTR.
   • BImplement a monitoring alert that triggers only when the CTR drop persists for a defined period (e.g., 48 hours) and exceeds a statistical threshold.
   • CRetrain the model immediately with recent data to adapt to the changed user behavior.
   • DIncrease the frequency of monitoring to collect more data points.
   Reveal Solution Discussion Select an answer first