Test Environments for AI-Based Systems CT-AI Practice Questions (Page 4)

Questions 16–20

1. 16foundation · easy

   What is the role of a test environment in a CI/CD pipeline for AI systems?

   • ATo store the final production model
   • BTo manually review all code changes
   • CTo automatically test model changes before they are deployed to production
   • DTo replace the need for version control
   Reveal Solution Discussion Select an answer first
2. 17foundation · easy

   What is the role of an API endpoint when serving an AI model in a test environment?

   • ATo store the training data
   • BTo manage user access to the test environment
   • CTo provide a standardized interface for sending requests to the model and receiving predictions
   • DTo visualize model performance
   Reveal Solution Discussion Select an answer first
4. 18foundation · easy

   Why is a dedicated test environment particularly important for AI-based systems compared to traditional software?

   • ATo allow the AI system to access production data without restrictions
   • BTo reduce the need for version control of training data
   • CTo isolate the AI system from production data and provide a safe, reproducible space for experimentation
   • DTo eliminate the need for monitoring during testing
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A bank is setting up a test environment for an AI model that uses customer transaction data. The data is subject to strict privacy regulations. The team needs to ensure that only the model testing team can access the data and that all access is logged. What is the most important security measure to implement?

   • AUse a shared test environment with no access restrictions to speed up development.
   • BEncrypt the data at rest and in transit, but allow all employees to access it.
   • CImplement role-based access control (RBAC) and enable audit logging for all data access.
   • DAnonymize the data and store it in a public cloud bucket.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A retail company is testing a recommendation model. The data science team updates the training dataset frequently and needs to ensure that test results can be traced back to the exact data used. They also need to restrict access to the test data to only the model evaluation team. What should they implement?

   • AUse only synthetic data to avoid the need for versioning.
   • BGrant all team members full access to the data to encourage collaboration.
   • CVersion control for the dataset with a unique identifier for each version, and access controls on the data storage.
   • DStore all data in a shared network drive without versioning.
   Reveal Solution Discussion Select an answer first