AI Ownership, Oversight, and Accountability AAIR Practice Questions (Page 2)

Questions 6–10

1. 6expert · hard

   A large hospital network uses an AI system to prioritize patient triage in emergency departments. The system was developed by a research institute and is used by clinicians. The hospital's risk committee is concerned about liability if the system mis-prioritizes a patient. The clinicians argue that they should not be held accountable for decisions made by an AI. Which approach best establishes clear accountability while addressing clinician concerns?

   • AAssign accountability to the research institute that developed the AI system
   • BEstablish a joint accountability committee of hospital and research institute staff
   • CMake each clinician individually accountable for every AI recommendation they follow
   • DDesignate the hospital's chief medical officer as the accountable owner, with clinicians responsible for appropriate use
   Reveal Solution Discussion Select an answer first
2. 7foundation · easy

   Why is it important to document AI roles and responsibilities?

   • ATo replace the need for training
   • BTo keep information confidential from stakeholders
   • CTo provide a clear reference for accountability and decision-making
   • DTo satisfy legal requirements only
   Reveal Solution Discussion Select an answer first
4. 8application · medium

   A bank is integrating AI into its loan approval process. The bank's ERM framework includes risk appetite statements for credit risk and operational risk. The AI system introduces model risk and potential bias. How should the bank ensure that AI risks are managed within its existing risk governance?

   • ACreate a separate AI risk framework that operates independently of the ERM framework
   • BIgnore AI-specific risks until they materialize, as they are unlikely to occur
   • CDelegate AI risk management to the data science team, as they understand the models
   • DIncorporate AI risks into the existing ERM framework by mapping them to relevant risk categories
   Reveal Solution Discussion Select an answer first
5. 9foundation · easy

   In an organization, which role is typically assigned ultimate responsibility for the outcomes of an AI system?

   • AThe executive or senior manager designated as the AI owner
   • BThe IT help desk technician
   • CThe data scientist who developed the model
   • DThe external vendor who supplied the AI software
   Reveal Solution Discussion Select an answer first
6. 10expert · hard

   A healthcare organization has implemented an AI governance framework, but recent incidents have revealed that some AI systems were deployed without proper oversight. The organization wants to improve its monitoring and review processes to prevent this in the future. What is the most effective action?

   • AImplement a mandatory pre-deployment review process for all AI systems, with sign-off from the accountable owner
   • BConduct a post-incident review after each incident to identify what went wrong
   • CIncrease the frequency of AI governance committee meetings
   • DRely on the AI governance committee to catch issues during periodic audits
   Reveal Solution Discussion Select an answer first