4.3 Describe the Importance of Responsible AI in Business. GENERATIVE-AI-LEADER Practice Questions (Page 2)

Questions 6–10

1. 6foundation · medium

   What is a key benefit of transparency in AI for building stakeholder trust?

   • AIt guarantees that the AI system will be free from bias.
   • BIt allows stakeholders to verify that the AI system is being used appropriately.
   • CIt eliminates the need for external audits of the AI system.
   • DIt ensures that the AI system's performance is always optimal.
   Reveal Solution Discussion Select an answer first
2. 7expert · hard

   A government agency is developing a generative AI system to summarize public feedback on proposed regulations. The training data includes a large number of submissions from a well-organized advocacy group and very few from other groups. The system now over-represents the advocacy group’s views in summaries. The agency must ensure balanced representation. What is the most effective action?

   • ACollect more feedback from underrepresented groups before training
   • BUse a larger language model to improve summarization quality
   • CPost-process the summaries to manually remove advocacy group language
   • DRe-sample the training data to balance the representation of different groups
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A hiring platform uses a generative AI tool to screen resumes. The training data is from the past five years and reflects the company’s historical hiring patterns, which favored candidates from certain universities. The tool now ranks candidates from those universities higher. What is the best way to mitigate this bias?

   • AIncrease the number of resumes processed per day
   • BRemove all university names from the resumes before processing
   • CAudit the training data and rebalance it to remove biased patterns
   • DUse a more complex model to improve accuracy
   Reveal Solution Discussion Select an answer first
5. 9foundation · medium

   What is re-identification in the context of AI privacy risks?

   • AThe process of training a model on the same data multiple times.
   • BThe process of linking anonymized data back to an individual.
   • CThe process of encrypting data in transit.
   • DThe process of deleting personal data from a dataset.
   Reveal Solution Discussion Select an answer first
6. 10application · medium

   A bank is training a generative AI model to summarize loan application documents. The training data is mostly from approved loans, with very few denied applications. The model now generates summaries that are overly optimistic about approval chances. What is the root cause of this issue?

   • AThe model lacks sufficient computational resources
   • BThe model has a bias due to imbalanced training data
   • CThe model is overfitting to the test set
   • DThe model is not using a large language model architecture
   Reveal Solution Discussion Select an answer first