Which of the following should be of GREATEST concern to an IS auditor when reviewing ethical considerations for an AI solution?
Answer(s): A
An unexplainable decision-making process is the greatest ethical concern because stakeholders cannot understand, justify, or challenge the AI’s outcomes, creating risks of unfair, biased, or unaccountable decisions.
Which of the following would pose the GREATEST risk when reviewing AI acceptable use training content?
Answer(s): B
Failing to address the non-deterministic nature of AI creates the greatest risk because users may incorrectly assume AI outputs are consistent, predictable, or authoritative. This misunderstanding can lead to overreliance on AI, improper validation of outputs, and significant governance and compliance issues.
An IS auditor is reviewing an AI application that uses customer data to refine the organization's marketing outreach strategies. Which of the following should be the auditor's PRIMARY focus during this review?
Answer(s): C
When customer data is used for AI-driven marketing, the primary risk is improper handling of personal data. Ensuring compliance with privacy regulations is essential to prevent unauthorized use, protect customer rights, and avoid legal or regulatory violations.
An organization is training a skin cancer recognition model. Photographs collected from which of the following sources would present the GREATEST risk associated with data integrity?
Images from a social media platform pose the greatest data integrity risk because they lack verified clinical labeling, consistent image quality, and controlled collection conditions. This greatly increases the likelihood of incorrect or misleading training data for a medical diagnostic model.
An IS auditor examining change management procedures for an AI system observes inconsistent training data validation and verification protocols prior to model retraining. Which of the following is the MOST significant risk in this context?
Inconsistent validation and verification of training data can allow corrupted, low-quality, or improperly prepared data into the retraining process, directly degrading the AI system’s reliability. This risk is more significant than documentation or timing issues because compromised data undermines the integrity and trustworthiness of the model’s outputs.
Post your Comments and Discuss ISACA AAIA exam prep with other Community members: