Principles of Ethical and Fair AI Use AIE Practice Questions (Page 1)

Questions 1–5

1. 1application · medium

   A financial institution is developing an AI system to approve small business loans. The model uses many features, including some that are not easily interpretable. The compliance team requires that any adverse action (loan denial) be explained to the applicant. Which approach best satisfies the accountability and transparency principles?

   • AProvide only the model's confidence score to the applicant as the explanation.
   • BAssign a human reviewer to personally call every denied applicant and explain the decision.
   • CUse a post-hoc explainability tool (e.g., SHAP) to generate feature-level explanations for each denial.
   • DReplace the model with a simple decision tree so every decision is inherently transparent.
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A recruitment firm uses an AI system to screen job applicants. The system was trained on resumes from the past decade, during which the firm predominantly hired male candidates for technical roles. The system now ranks male candidates higher. The firm wants to address this bias. Which action is most appropriate?

   • AIncrease the model's complexity to better capture subtle patterns.
   • BAudit the training data for bias and use techniques such as re-weighting or synthetic data generation to balance the dataset.
   • CBlindly remove gender-related features from the model.
   • DUse the model only for initial screening and rely on human recruiters for final decisions.
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A city government deploys an AI system to prioritize road maintenance projects. The system uses a complex ensemble model that is difficult to interpret. Residents and city council members request an explanation of why certain roads are prioritized over others. Which approach best addresses the transparency and explainability requirements?

   • AProvide a detailed technical report of the model's architecture and training data to the public.
   • BUse a simpler, interpretable model such as a decision tree or logistic regression, even if it slightly reduces predictive accuracy.
   • CPublish only the final priority list without any explanation.
   • DKeep the model's logic confidential to protect the city's competitive advantage.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A company is developing an AI system that uses facial recognition to identify customers entering its stores. The system will be used to personalize offers. The company wants to assess the ethical impact on customers and employees. Which stakeholder impact should be prioritized?

   • AThe potential for increased sales from personalized offers.
   • BThe accuracy of the facial recognition system in identifying customers.
   • CThe privacy and consent of customers and employees who are scanned.
   • DThe cost of implementing the facial recognition system.
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   What does accountability in AI development refer to?

   • AAssigning responsibility for the outcomes and impacts of AI systems
   • BMaking AI systems fully autonomous
   • CEnsuring that AI systems are fast and efficient
   • DKeeping AI algorithms secret to protect intellectual property
   Reveal Solution Discussion Select an answer first