Change Management Specific to AI AAIA Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   What is the primary purpose of monitoring an AI system after a change?

   • ATo delete old logs
   • BTo automatically revert the change after a fixed time
   • CTo train the model on new data
   • DTo ensure the system remains stable and performs as expected
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A team is updating a computer vision model used for quality control in manufacturing. The new model has higher accuracy on the test set, but it is slower and requires more computational resources. The production environment has limited GPU capacity. What should the team do before deployment?

   • APerform a performance test to ensure the model meets latency and resource requirements in production
   • BOnly test the model's accuracy, not its performance
   • CDeploy the model and hope the GPUs can handle the load
   • DRoll back the model before testing
   Reveal Solution Discussion Select an answer first
4. 8foundation · easy

   What type of testing is used to evaluate the performance of an AI model on unseen data?

   • AUnit testing
   • BIntegration testing
   • CValidation testing
   • DSmoke testing
   Reveal Solution Discussion Select an answer first
5. 9application · medium

   A hospital is deploying an AI system that predicts patient readmission risk. The change management process must consider the impact on clinicians' workflow and patient outcomes. Which action best addresses the unique challenges of AI change management in this context?

   • AConduct a multidisciplinary impact assessment that includes clinical, technical, and ethical perspectives
   • BSchedule a post-change review to evaluate model accuracy
   • CObtain approval from the hospital's IT change advisory board
   • DEnsure the model is trained on the most recent patient data available
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   A data science team is developing a new model version. They have updated the training data and the model architecture. They want to ensure that the exact combination of data, code, and model can be reproduced for audit purposes. What is the best practice?

   • AStore the model file in a shared drive and document the changes in a wiki
   • BCreate a versioned release that includes the dataset, code, and model artifact with a unique identifier
   • COnly version the code, assuming the data is static
   • DUse a random naming convention for model files
   Reveal Solution Discussion Select an answer first