Selecting Test Techniques for AI-Based Systems CT-AI Practice Questions (Page 3)

Questions 11–15

1. 11foundation · easy

   When selecting test techniques for an AI system, why is it important to consider regulatory requirements?

   • ARegulatory requirements determine the programming language to use
   • BRegulatory requirements may mandate specific testing activities to ensure compliance
   • CRegulatory requirements are only relevant for non-AI systems
   • DRegulatory requirements have no impact on testing
   Reveal Solution Discussion Select an answer first
2. 12application · medium

   A healthcare AI system is being tested for patient risk prediction. The system must comply with regulations that require fairness across demographic groups. The team has access to the training data and the model. Which testing approach should they prioritize?

   • ASecurity testing for adversarial attacks
   • BFairness testing using subgroup analysis and equalized odds
   • CPerformance testing with a large test set
   • DUsability testing with clinicians
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which technique is used to assess how well an AI model performs on unseen or adversarial data?

   • AIntegration testing
   • BUnit testing
   • CRegression testing
   • DRobustness testing
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   Which AI-specific test technique compares the outputs of multiple versions of a system to identify discrepancies?

   • AMetamorphic testing
   • BExploratory testing
   • CDifferential testing
   • DMutation testing
   Reveal Solution Discussion Select an answer first
6. 15application · medium

   A team is testing an AI system that classifies medical images. The model is highly complex, and the team has access to the model's architecture and weights. They need to assess the model's internal decision-making process. Which testing approach is most suitable?

   • ABlack-box testing with a large set of labeled images
   • BDifferential testing against a simpler model
   • CMetamorphic testing with image transformations
   • DWhite-box testing using neuron coverage and gradient-based analysis
   Reveal Solution Discussion Select an answer first