Principles of Ethical and Fair AI Use AIE Practice Questions (Page 4)

Questions 16–20

1. 16application · medium

   A manufacturing company plans to use AI to optimize energy consumption in its factories. The system will reduce costs and carbon emissions but may lead to job losses for some workers. The company wants to assess the ethical impact. Which stakeholder group should be given the most consideration?

   • AWorkers who may be displaced by automation.
   • BShareholders who will benefit from reduced costs.
   • CCustomers who may see lower prices.
   • DThe environment, which will benefit from reduced emissions.
   Reveal Solution Discussion Select an answer first
2. 17application · medium

   A nonprofit organization uses an AI system to allocate resources for disaster relief. The system prioritizes areas based on predicted need, but some regions with less data may be underserved. The organization wants to ensure ethical decision-making. Which approach is most appropriate?

   • ARely solely on the AI system to avoid human bias.
   • BUse the AI system only for areas with sufficient data.
   • CCombine the AI system's recommendations with human oversight and local knowledge.
   • DKeep the AI system's methodology secret to avoid criticism.
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A company is developing an AI system to recommend which job applicants to interview. The system was trained on data from the past five years, during which the company hired mostly from certain universities. The company wants to increase diversity. What is the most appropriate action?

   • ARely on the HR team to manually override the system's recommendations when they suspect bias.
   • BRemove university names from the model inputs to ensure fairness.
   • CUse the system as-is because it reflects the company's successful hiring patterns.
   • DAudit the system to see if it ranks applicants from certain universities higher, and adjust the model to reduce that bias.
   Reveal Solution Discussion Select an answer first
5. 19foundation · easy

   Which of the following is a common source of bias in AI systems?

   • ABiased training data
   • BSmall model size
   • CHigh model accuracy
   • DUsing too many features
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A bank deploys a credit-scoring model that uses an applicant's postal code as a feature. The model's output shows significantly lower approval rates for applicants in certain neighborhoods. The bank's compliance team asks the data science team to evaluate the model for fairness before it goes into production. Which action best addresses the fairness concern while preserving the model's utility?

   • AAdd more features from the same neighborhoods to increase the model's accuracy for those groups.
   • BApply a fairness metric such as demographic parity or equalized odds and adjust the decision threshold to meet the chosen metric.
   • CUse a more complex model, such as a deep neural network, to better capture the relationship between postal code and creditworthiness.
   • DRemove the postal code feature and retrain the model without it.
   Reveal Solution Discussion Select an answer first