AAISM practice questions 71–80

1. Question 71 · 82 in bank

   A bank's fraud detection AI model is trained on customer transaction data. A customer submits a data erasure request under privacy law, but the data also exists in training sets. Which of the following is the BEST approach if there is concern that complete data removal may reduce accuracy?

   • ADeny the request because legitimate purpose of use outweighs privacy rights.
   • BSuspend the AI model until all erasure requests have been processed. B. Update standard operating procedures to ensure that the customer's data is fully deleted.
   • CInform the customer that deletion from AI models is technically infeasible.

   Show answer

   Answer: C

   Next question

2. Question 72 · 82 in bank

   Which of the following is the MOST important reason to establish AI-specific key risk indicators (KRIs)?

   • AExplainability and transparency deficiencies can be more quickly identified.
   • BAI systems learn and change as they interact with dynamic environments.
   • CContinuous availability is more critical for AI systems than for traditional systems.
   • DKRIs help to identify potential threats that may exploit AI system vulnerabilities.

   Show answer

   Answer: B

   Next question

3. Question 73 · 82 in bank

   When developing organizational generative AI policies, the FIRST step is to:

   • Aunderstand how the organization intends to ethically use generative AI.
   • Badopt current AI legal and regulatory requirements.
   • Cidentify potential risk associated with the deployment of generative AI.
   • Dinventory the current state of AI systems, use cases, and processes.

   Show answer

   Answer: D

   Next question

4. Question 74 · 82 in bank

   Which of the following is the BEST approach for improving the robustness of an AI system in production?

   • AReducing model complexity to simplify inference and minimize computational errors
   • BEncrypting training datasets to protect against data theft
   • CLimiting user access to the AI system to prevent misuse
   • DImplementing automated retraining to handle drift and performance degradation

   Show answer

   Answer: D

   Next question

5. Question 75 · 82 in bank

   Which of the following processes involves tagging and labeling raw datasets to create training data sets?

   • ACategorization
   • BTransformation
   • CAnnotation
   • DClassification

   Show answer

   Answer: C

   Next question

6. Question 76 · 82 in bank

   An organization is implementing a large language model (LLM) to answer questions for both employees and customers. The CIO wants the LLM to provide employees with additional internal data that should not be shared with customers. Which of the following is the MOST secure approach?

   • ARetrieval-augmented generation (RAG) with access control lists (ACLs)
   • BModel fine tuning with increased use of function calls
   • CPrompt sanitization
   • DAllow list employee access

   Show answer

   Answer: A

   Next question

7. Question 77 · 82 in bank

   A bank is training an AI credit scoring model on a vast amount of sensitive customer data. Which of the following is the MOST critical security risk in this training process?

   • AData poisoning to corrupt training outcomes
   • BAlgorithmic bias leading to discriminatory lending
   • CResources needed to retrain and validate the AI model
   • DModel inversion or membership inference attacks

   Show answer

   Answer: D

   Next question

8. Question 78 · 82 in bank

   A pharmaceutical organization uses an AI solution from a third-party vendor to analyze confidential research and development data. Which of the following controls is MOST essential to ensure the vendor's AI solution complies with the organization's security requirements?

   • ACommunicate security requirements to the vendor.
   • BMinimize sensitive data being used by the vendor's AI solution.
   • CImplement a private instance of the solution.
   • DReview the vendor's independent security audit report.

   Show answer

   Answer: D

   Next question

9. Question 79 · 82 in bank

   Which of the following would BEST ensure safe implementation of AI systems?

   • ASelection of AI projects that align with business objectives.
   • BCompliance with the AI risk management framework and applicable regulations.
   • CEstablishment of an AI program steering committee to govern AI implementation.
   • DPhased implementation with human-in-the-loop validation before full rollout.

   Show answer

   Answer: D

   Next question

10. Question 80 · 82 in bank

    To facilitate the adoption of a new AI system, an organization has provided data access to third-party solutions. Which of the following BEST mitigates risk associated with third-party software?

    • AUpdating tolerance levels to accept new risk
    • BDesignating boundaries for AI operation
    • CConsulting AI experts to help mitigate the risk
    • DConducting regular external audits

    Show answer

    Answer: B

    Next question

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

AAISM practice, question 71. 82 questions available.