AI Implementation, Maintenance, and Decommissioning AAIR Practice Questions (Page 2)

Questions 6–10

1. 6application · medium

   A retail company is planning to implement an AI system for inventory demand forecasting. The implementation plan must include a validation strategy that ensures the model performs well across different product categories and seasons. The company also wants to plan for the eventual decommissioning of the system. Which element should be included in the implementation plan?

   • AInclude a decommissioning plan that specifies data retention periods for all historical sales data.
   • BUse a simple validation approach that only checks overall accuracy.
   • CSchedule a single validation test at the end of the implementation phase.
   • DDefine validation metrics and acceptance criteria for each product category and season.
   Reveal Solution Discussion Select an answer first
2. 7application · medium

   A hospital is decommissioning an AI system that assisted with patient diagnosis. The system contains sensitive patient data. The hospital must ensure that the decommissioning complies with health data regulations and that patient data is handled appropriately. What should be the first step in the decommissioning plan?

   • AShut down the system and then review the data.
   • BConduct a data inventory to identify where patient data is stored and how it is used.
   • CNotify patients that their data will be deleted.
   • DDelete all patient data immediately to ensure privacy.
   Reveal Solution Discussion Select an answer first
4. 8foundation · easy

   What should be included in a decommissioning plan for an AI system?

   • AA plan to increase the model's training data
   • BA list of new features to add to the system
   • CA data retention and disposal strategy
   • DA strategy to improve model accuracy
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   Which of the following is a key consideration when planning the implementation of an AI system in an organization?

   • ASelecting the AI algorithm before defining the business problem
   • BGuaranteeing the model will never produce a false positive
   • CEnsuring the AI system aligns with the organization's risk appetite and governance framework
   • DEliminating the need for human oversight entirely
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   A multinational corporation is implementing an AI system for employee attrition prediction. The system will be deployed in multiple regions with different data protection regulations. The implementation plan must address the need for regular model updates and the ability to decommission the system in any region independently. The company has a centralized data science team but regional IT teams manage local infrastructure. Which implementation strategy best balances these requirements?

   • AUse a hybrid approach where the model is deployed centrally but all data is stored in a single global data lake.
   • BDeploy separate models for each region, each trained and maintained by the regional IT teams.
   • CDeploy a single model in the company's headquarters region and stream data from all regions to that location.
   • DDeploy a single centralized model with a global API, and use regional data stores that comply with local regulations.
   Reveal Solution Discussion Select an answer first