Objective 1.6 Identify Potential Ethical Concerns AIP-210 Practice Questions (Page 3)

Questions 11–12

1. 11expert · hard

   A credit union is building an AI to detect fraud. The model has high precision but low recall, meaning it misses many fraudulent transactions. The credit union is concerned about both financial losses and customer inconvenience from false positives. Which approach best balances these concerns?

   • AIncrease recall by flagging more transactions as fraudulent
   • BKeep the current high-precision model to minimize customer inconvenience
   • CAdjust the decision threshold to achieve a target trade-off between recall and precision based on the cost of each error type
   • DUse a different model that is less accurate but more balanced
   Reveal Solution Discussion Select an answer first
2. 12expert · hard

   A company is developing an AI to screen job applicants. The model is trained on resumes from a period when the company had a diversity problem. The company wants to improve diversity but also maintain hiring standards. Which approach best addresses this tension?

   • AUse the model as-is because it reflects the company's historical standards
   • BAdd a diversity quota to the model's output without changing the model
   • CRemove the model and rely on unstructured human review
   • DRetrain the model on a dataset that is balanced across demographic groups and validate that it does not penalize qualified candidates from underrepresented groups
   Reveal Solution Discussion Select an answer first