AI Risk Management AAIA Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A utility company uses AI for demand forecasting. The risk team is integrating AI risk into the ERM framework. The company's ERM framework uses a five-point scale for likelihood and impact. The AI risk team has been using a different scale. Which action best facilitates integration?

   • AUse only qualitative assessments for AI risks to avoid scale mismatches.
   • BMap the AI risk scale to the ERM scale and document the mapping in the risk register.
   • CContinue using the AI-specific scale, as it is more detailed and provides better risk differentiation.
   • DReport AI risks separately to the board, using the AI-specific scale, to preserve detail.
   Reveal Solution Discussion Select an answer first
2. 2foundation · easy

   Which of the following is a key component of an effective AI risk monitoring process?

   • ADefining key risk indicators (KRIs) that are monitored on an ongoing basis
   • BDeleting audit logs after each model update
   • CLimiting risk reporting to senior management only
   • DConducting a one-time risk assessment at the start of the project
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   A pharmaceutical company uses AI in drug discovery. The AI model has produced promising results, but there are concerns about the interpretability of its predictions for regulatory approval. The company must decide on a governance approach. Which approach best balances innovation and regulatory risk?

   • AAllow the R&D team to proceed with the AI model without additional governance, as long as it meets performance targets.
   • BOutsource all AI model validation to an external lab and rely on their certification for regulatory approval.
   • CStop all AI research until interpretability methods are fully mature, to avoid regulatory risk.
   • DEstablish a cross-functional AI governance board that includes R&D, legal, compliance, and regulatory affairs, with a mandate to review AI models for interpretability and regulatory compliance before approval.
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A multinational corporation is implementing AI across various departments. The board wants to ensure clear accountability for AI risk management. Which governance structure best achieves this?

   • AEstablish an AI risk committee with representatives from legal, compliance, IT, and business units, chaired by the Chief Risk Officer, with defined terms of reference and escalation paths.
   • BDelegate all AI risk decisions to the IT security team, as they understand the technology.
   • CCreate a data science center of excellence that reviews all AI projects for technical accuracy, with no formal risk reporting.
   • DAssign each department head to manage AI risks independently, reporting annually to the board.
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   A financial services firm is assessing the risk of an AI trading algorithm. The algorithm has a history of rare but severe losses. The risk team needs to assess the likelihood and impact. Which method is most appropriate?

   • AQualitatively rate the risk as 'low' because the algorithm has been profitable overall.
   • BAccept the risk without formal assessment because the algorithm is proprietary and well-tested.
   • CUse scenario analysis to model extreme market conditions and estimate potential losses, combined with historical loss data.
   • DRely solely on the average daily loss from the past year, as it is a simple quantitative measure.
   Reveal Solution Discussion Select an answer first