AI Design, Development/Procurement, and Documentation AAIR Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A bank is procuring an AI-based anti-money laundering (AML) system. The vendor offers two options: a cloud-based SaaS solution with lower upfront cost but data hosted in a foreign jurisdiction, and an on-premises solution with higher upfront cost but full data control. The bank's compliance team requires that customer data remain within the home country. The procurement team must balance cost and compliance. What is the MOST appropriate decision?

   • AChoose the on-premises solution because it ensures data residency compliance, even though it is more expensive.
   • BChoose the cloud-based SaaS solution and rely on the vendor's contractual promise of data residency.
   • CChoose the cloud-based SaaS solution because it is cheaper and assume the data residency requirement can be waived.
   • DDefer the decision until the vendor can provide a hybrid solution.
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   An AI team is developing a predictive model for patient readmission risk. The team has completed the design phase and is about to begin development. The project sponsor wants to ensure that the model is validated against clinical outcomes and that the validation process is documented for regulatory purposes. What is the MOST appropriate sequence of activities?

   • ADevelop the model and deploy it, then validate it in a live environment.
   • BDocument the validation plan before development, then develop and validate the model.
   • CDevelop the model, then validate it against clinical outcomes, and document the validation results.
   • DValidate the model against clinical outcomes before development to ensure it meets requirements.
   Reveal Solution Discussion Select an answer first
4. 3expert · hard

   An AI system used for credit decisions is being audited. The auditor requests documentation that shows the rationale for the model's design choices and how the system was tested for fairness. The organization has a design specification and a testing report, but the auditor also wants to see the decision-making process behind the design choices. What should the organization provide?

   • AThe resumes of the data scientists who designed the model.
   • BThe design specification and testing report, as they are sufficient.
   • CA decision log that records the rationale for design choices and the alternatives considered.
   • DA summary of the model's performance metrics and accuracy.
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   Which characteristic of documentation is most important for supporting regulatory compliance and auditability?

   • AConfidentiality
   • BVisual appeal
   • CBrevity
   • DTraceability
   Reveal Solution Discussion Select an answer first
6. 5application · medium

   An AI system used for medical diagnosis is being audited by a regulatory body. The auditor asks for evidence that the system was validated against clinical standards and that any changes were properly reviewed. What documentation should the organization provide?

   • AA validation report and a change management log.
   • BThe vendor's marketing materials and product brochures.
   • CA project timeline and budget report.
   • DA list of the system's users and their training records.
   Reveal Solution Discussion Select an answer first