Transparency CRAGE Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A company has developed an AI system that predicts employee attrition. The system is used by HR to identify at-risk employees. The company wants to be transparent with employees about the system's use, but is concerned that full transparency may cause anxiety or gaming of the system. What is the best approach?

   • ATell employees only that the system exists, but not what data it uses
   • BPublish the full model and all data used to be fully transparent
   • CInform employees that the system is used, explain its general purpose and data inputs, and provide a channel for questions or concerns
   • DKeep the system completely secret to avoid anxiety
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A city government deploys an AI system to prioritize road maintenance based on predicted infrastructure degradation. The city council wants to communicate the system's decisions to the public. Which communication strategy best balances transparency with public understanding?

   • ARelease a plain-language summary of how the system works, its limitations, and the factors that influence prioritization
   • BHold a press conference with the data science team to explain the model's mathematical details
   • CPublish the raw sensor data and model code on an open data portal
   • DProvide a detailed technical report to each citizen who requests it
   Reveal Solution Discussion Select an answer first
4. 3foundation · easy

   Why is it important to communicate the limitations of an AI system to stakeholders?

   • ATo protect the organization from legal liability in all cases
   • BTo ensure stakeholders have realistic expectations and can make informed decisions about relying on the AI's outputs
   • CTo discourage stakeholders from using the AI system
   • DTo demonstrate the technical complexity of the AI system
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   An insurance company uses a machine learning model to set premiums. Regulators require the company to explain why a specific customer received a high premium. The model is a black-box ensemble. Which combination of mechanisms would best enable the company to meet this regulatory request?

   • AComprehensive logging of all training data and a one-time model audit
   • BA detailed data dictionary and a user-friendly FAQ page
   • CAn interpretable linear model and a public disclosure of all model coefficients
   • DModel documentation and post-hoc explainability tools like SHAP or LIME
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A company's AI system uses a complex neural network to approve or reject equipment warranty claims. The customer service team needs to explain rejections to customers. Which transparency mechanism would best support the customer service team?

   • AA real-time explanation tool that generates a simple reason for each rejection
   • BThe model's training dataset and code repository
   • CA static FAQ document describing the warranty claim process
   • DA monthly report summarizing the model's overall approval rate
   Reveal Solution Discussion Select an answer first