Change Management Specific to AI AAIA Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   What is a common trigger for retraining an AI model in production?

   • AA change in the color scheme of the user interface
   • BA drop in model performance or detection of data drift
   • CA scheduled calendar date
   • DAn increase in the number of users
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   An organization is required to provide evidence of model versioning for a regulatory audit. The audit requires that the exact model, data, and code used in production at a specific date be reproducible. What is the best way to meet this requirement?

   • AKeep a log of model changes in a spreadsheet
   • BStore only the final model file for each version
   • CRely on the data science team's memory to reconstruct the version
   • DUse a version control system that tracks all assets and provides a snapshot of the production environment
   Reveal Solution Discussion Select an answer first
4. 13application · medium

   A team is updating a natural language processing model used for customer support. They have a new model that performs better on the test set. What testing strategy is essential before deploying to production?

   • ARun regression tests on a set of known edge cases and compare performance with the current model
   • BDeploy the model to production and monitor its performance for a week
   • CTest the model only on the new test set that was used during development
   • DAsk the customer support team to manually review all model outputs for a day
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   Which of the following is a stage in the AI model lifecycle where changes to the model itself (e.g., architecture or hyperparameters) are made?

   • ADeployment only
   • BModel training and tuning
   • CData collection
   • DMonitoring only
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   What is the primary purpose of a change approval board (CAB) in AI change governance?

   • ATo monitor the AI system after deployment
   • BTo retrain the AI model
   • CTo review and authorize AI changes before they are implemented
   • DTo write the code for AI changes
   Reveal Solution Discussion Select an answer first