AIF-C01 Practice Question 59: Undisclosed AI Overreliance

The Responsible AI Lens frames exactly this scenario as a transparency problem: it tells you to *consider situations where system documentation is insufficient, users do not understand the probabilistic nature of a system output, or where users are unaware of AI system presence*. It asks teams to *create simple tests that check whether users understand when they're interacting with AI* and warns that transparency gaps *might lead to harmful over-reliance on uncertain outputs*, which is precisely what the review found (option C).

- A is wrong because controllability is about the operators' ability to steer, monitor and stop the system. Nothing in the finding says the bank has lost control of the assistant; the gap is in what the *customer* knows. - B is wrong because no data was exposed or mishandled. The customers' information was fine; their understanding of what they were talking to was not. - D is wrong because governance covers the policies, ownership and review processes that sit around an AI system. Weak governance may well be why the disclosure was never designed, but the defect the review identified is the missing disclosure itself.

Reference: RAIRC03-BP10 Measure transparency quality