A major insurer uses an AI model to detect claims fraud. The model's recall for actual fraudulent claims is 72%, meaning 28% of real fraud goes undetected. The insurer's leadership argues this is acceptable because 'no model is perfect.' What should the AI risk manager consider when evaluating this performance level?

B is correct because risk management of a fraud detection model requires evaluating three dimensions: the materiality of the missed fraud (financial exposure from 28% undetected fraud), whether model improvements are feasible, and whether the miss rate is consistent across demographics (some groups being missed at higher rates would be a fairness issue). Simply accepting 72% recall as 'good enough' (A) without this analysis is insufficient. Targeting 100% recall (C) would cause operationally unacceptable false positives. Regulatory reporting (D) may be appropriate for material known issues but is not the first step.