Transparency CRAGE Practice Questions (Page 2)

Questions 6–10

1. 6foundation · easy

   Which of the following is an example of a transparency mechanism in AI governance?

   • AUsing a more complex neural network to improve accuracy
   • BEncrypting all training data to protect privacy
   • CRestricting access to the AI system to only a few authorized users
   • DMaintaining detailed documentation of the model's design, data, and decision-making process
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A government agency is considering two AI systems for predicting recidivism. System X is a simple logistic regression that is fully interpretable but has lower accuracy. System Y is a complex ensemble that is more accurate but requires post-hoc explanations. The agency must be transparent to the public and comply with regulations. Which system should the agency choose?

   • AChoose System X because interpretability is more important than accuracy in high-stakes decisions
   • BChoose System Y and invest in robust explainability tools and documentation
   • CChoose System Y but do not disclose its use to the public to avoid scrutiny
   • DChoose System X and provide no further explanation because it is interpretable
   Reveal Solution Discussion Select an answer first
4. 8expert · hard

   A credit bureau uses a complex gradient boosting model for credit scoring. Regulators demand explanations for individual credit decisions. The model is retrained monthly, and the bureau must provide explanations for decisions made in the previous month. What is the most effective approach?

   • AUse a model-agnostic explainability tool (e.g., LIME or SHAP) on the specific model version that was in production for each decision
   • BUse the current model version to explain all past decisions
   • CSwitch to a simple linear model to make explanations easier, even if accuracy drops
   • DProvide only the input features and output score without any explanation
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   When communicating about an AI system to a non-technical audience, what is the most effective approach?

   • AUse technical jargon and detailed algorithmic formulas to provide complete information
   • BProvide high-level summaries that focus on capabilities, limitations, and practical implications in plain language
   • CShare only the positive aspects of the AI system to avoid causing concern
   • DAvoid discussing the AI system altogether to prevent misunderstandings
   Reveal Solution Discussion Select an answer first
6. 10foundation · easy

   Which of the following is an example of a regulatory framework that mandates transparency in AI systems?

   • AThe Health Insurance Portability and Accountability Act (HIPAA) in the United States
   • BThe General Data Protection Regulation (GDPR) in the European Union, which includes provisions for the right to explanation in certain automated decisions
   • CThe Payment Card Industry Data Security Standard (PCI DSS)
   • DThe International Organization for Standardization (ISO) 9001 quality management standard
   Reveal Solution Discussion Select an answer first