Question 22
Domain 2 — AI Operations, Lifecycle, and Control EnvironmentWhich of the following is the PRIMARY reason IS auditors must be aware that generative AI may return different investment recommendations from the same set of data?
Correct answer: C
Explanation
Generative AI does not produce fixed outputs; it generates responses using probabilistic models, so the same data can lead to different recommendations. Because its computational logic is based on probabilities, IS auditors must expect variability and assess the reliability of any investment advice it produces.
Why each option is right or wrong
A. Limitations can arise in the quantification of risk profiles.
B. Neural node access varies each time the process is executed.
C. Computational logic is based on probabilities.
Generative AI systems do not execute deterministic rules; they generate outputs by sampling from probabilistic models, so the same input set can legitimately produce different recommendations on separate runs. For an IS auditor, the key issue is that the model’s logic is not fixed or repeatable in the way a traditional programmed decision engine is, which means output variability is inherent rather than anomalous.
D. Servers are reconfigured periodically.