AI Implementation, Maintenance, and Decommissioning AAIR Practice Questions (Page 3)

Questions 11–15

1. 11application · medium

   An online media company has deployed an AI system that personalizes content recommendations. The system's performance is monitored using click-through rate (CTR). After a major algorithm update, the CTR drops significantly. The company needs to determine whether the drop is due to the model update or external factors. What should they do?

   • AReview the model's training data to see if it is outdated.
   • BIncrease the frequency of monitoring to gather more data.
   • CCompare the CTR of the new model against a control group that still uses the old model.
   • DRoll back to the previous model immediately.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   What is a primary objective when deploying an AI model into a production environment?

   • AIsolating the model from all other organizational processes
   • BRemoving all logging and monitoring capabilities
   • CMaximizing the number of model parameters
   • DIntegrating the model with existing systems and data flows
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Why is stakeholder communication important when planning the decommissioning of an AI system?

   • ATo increase the model's performance before shutdown
   • BTo advertise the system to new users
   • CTo ensure all affected parties are aware of the timeline and impact
   • DTo obtain approval for additional funding
   Reveal Solution Discussion Select an answer first
5. 14application · medium

   A bank is implementing an AI model for credit risk assessment. The model must comply with regulatory requirements for explainability. The bank's compliance team requires that the model's decisions can be explained to customers. Which validation activity should be included?

   • ATest the model with a set of explainability tools (e.g., SHAP, LIME) to generate feature importance explanations for individual predictions.
   • BConduct a user study to see if customers understand the model's decisions.
   • CEnsure the model uses only a small number of features to make it inherently interpretable.
   • DValidate the model's overall accuracy on a test set.
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A company is decommissioning an AI system that was used for customer churn prediction. The system is integrated with the CRM and a data warehouse. The company wants to ensure that the decommissioning does not disrupt other systems that may rely on the data. What should be done before shutting down the AI system?

   • AMap all data flows and dependencies to identify which downstream systems consume the AI system's outputs.
   • BDelete all data from the data warehouse to ensure no residual data remains.
   • CShut down the system immediately and then assess the impact on other systems.
   • DKeep the AI system running in a degraded mode until all other systems are updated.
   Reveal Solution Discussion Select an answer first