AI Models, Frameworks, Strategies, and Use Cases AAIR Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which AI-specific risk is most directly associated with a model producing different outcomes for similar individuals based on protected characteristics?

   • ARobustness
   • BPrivacy
   • CExplainability
   • DBias
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   An agricultural company is using a machine learning model to predict crop yields based on weather and soil data. The model performs well in the region where it was trained, but the company wants to expand to a new region with different climate patterns. What is the most important governance consideration?

   • AThe model's training data may contain personal data of farmers
   • BThe model's explainability may be insufficient for regulators in the new region
   • CThe model's performance may degrade in the new region due to distribution shift
   • DThe model's framework may not be compatible with the new region's data formats
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A city government is using AI to optimize traffic light timings to reduce congestion. The system uses reinforcement learning to adjust signals in real time. What is a key governance consideration for this use case?

   • AThe system may prioritize certain roads over others, leading to inequitable traffic flow
   • BThe system's framework may not be compatible with existing traffic infrastructure
   • CThe system's model may be too complex to explain to the public
   • DThe system's training data may contain personal information about drivers
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   A retail company uses AI to personalize product recommendations. What is a key benefit and a key risk of this use case?

   • ABenefit: guaranteed customer satisfaction; Risk: higher operational costs
   • BBenefit: elimination of all marketing costs; Risk: model bias
   • CBenefit: reduced need for data collection; Risk: lower sales
   • DBenefit: increased customer engagement; Risk: potential privacy concerns from collecting user data
   Reveal Solution Discussion Select an answer first
6. 25application · medium

   A company has developed a machine learning model for fraud detection. Before deployment, they want to ensure the model meets regulatory requirements for fairness and transparency. Which stage of the AI model lifecycle should they focus on?

   • ADesign stage, to define the problem and success metrics
   • BDeployment stage, to monitor the model in production
   • CValidation stage, to test the model against fairness and transparency criteria
   • DDevelopment stage, to train the model with diverse data
   Reveal Solution Discussion Select an answer first