Integrating AI into Enterprise Systems CAIPM Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which governance framework component is essential for ensuring AI systems comply with regulations during integration?

   • AUsing only proprietary algorithms to avoid open-source licensing issues
   • BDeleting all audit logs to protect user privacy
   • CAllowing AI systems to make decisions without human review
   • DEstablishing clear accountability and oversight for AI decisions
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A healthcare organization is integrating an AI system that triages patient messages. The system will process protected health information (PHI). The organization must comply with HIPAA and wants to use a public cloud AI service. What should the program manager ensure before approving the integration?

   • AThe AI service is deployed only in the organization's on-premises data center.
   • BThe cloud provider offers a HIPAA-eligible service and the organization signs a Business Associate Agreement (BAA).
   • CAll PHI is anonymized before being sent to the AI service, so no BAA is needed.
   • DThe AI service is accessed through a private endpoint and no data leaves the organization's network.
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A healthcare provider is integrating an AI model that analyzes medical images. The model is deployed as a containerized service that receives images from the hospital's imaging system and returns diagnostic suggestions. The hospital's network is isolated from the internet for security. What should the architecture include?

   • AExpose the AI service on the hospital's internal network and have the imaging system call it via a private API endpoint.
   • BDeploy the AI model on the same server as the imaging system to avoid network calls.
   • CUse a USB drive to transfer images to a separate computer that runs the AI model, then manually enter results into the imaging system.
   • DUpload images to a public cloud AI service and download the results back to the hospital.
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   What is the main objective of monitoring an AI system after deployment?

   • ATo replace the AI model with a new one every week
   • BTo reduce the number of users accessing the system
   • CTo increase the complexity of the system architecture
   • DTo detect performance degradation, data drift, and system failures in real time
   Reveal Solution Discussion Select an answer first
6. 25expert · hard

   A bank is integrating an AI model for real-time fraud detection. The model needs to access transaction data from a legacy mainframe that only supports nightly batch exports. The bank's fraud team needs real-time alerts to stop fraudulent transactions. The mainframe cannot be modified in the near term. What should the integration strategy include?

   • AReplace the mainframe with a modern system that supports real-time APIs.
   • BUse a screen-scraping tool to read transaction data from the mainframe's user interface in real time.
   • CRely solely on the nightly batch data and accept that fraud alerts will be delayed by up to 24 hours.
   • DUse the nightly batch data to train the model, but also implement a real-time event stream from the mainframe's transaction logs using a change data capture (CDC) tool.
   Reveal Solution Discussion Select an answer first