Testing Techniques for AI Solutions AAIA Practice Questions (Page 7)

Questions 31–35

1. 31application · medium

   An e-commerce company's AI-based recommendation engine is expected to handle a 10x increase in traffic during the holiday season. The system currently responds within 200 ms under normal load. The team wants to ensure the system remains responsive and does not degrade the user experience during peak traffic. Which testing approach should they use?

   • AReview the model's training data for biases that could affect recommendation quality.
   • BDeploy the system to production and monitor its performance during the holiday season.
   • CRun a load test simulating the expected peak traffic and measure response times, throughput, and resource utilization.
   • DTest the system with a small set of typical user queries to verify the recommendations are accurate.
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   Why is reproducibility important in model validation?

   • AIt ensures the model is free of security vulnerabilities
   • BIt ensures the model is fair across demographic groups
   • CIt ensures that the same evaluation yields consistent results, confirming the reliability of performance metrics
   • DIt ensures the model can handle invalid inputs
   Reveal Solution Discussion Select an answer first
4. 33foundation · easy

   What is the focus of robustness testing for an AI solution?

   • AVerifying that the solution meets functional requirements
   • BIdentifying security vulnerabilities such as data poisoning
   • CMeasuring the solution's performance under peak load
   • DEvaluating the solution's ability to handle invalid, unexpected, or adversarial inputs without failure
   Reveal Solution Discussion Select an answer first
5. 34foundation · easy

   What is the primary objective of security testing for an AI solution?

   • ATo evaluate the AI solution's ability to handle invalid inputs
   • BTo identify and mitigate security vulnerabilities, including data poisoning, model inversion, and adversarial attacks
   • CTo verify that the AI solution meets functional requirements
   • DTo measure the AI solution's performance under peak load
   Reveal Solution Discussion Select an answer first
6. 35application · medium

   A hospital is using an AI system to recommend treatment plans for patients. Doctors need to understand why the system makes certain recommendations, and the hospital must document these explanations for medical review boards. Which testing activity best supports this requirement?

   • AExplainability and interpretability testing to generate human-understandable explanations for the model's decisions.
   • BPerformance testing to ensure the system responds quickly during consultations.
   • CRegression testing to ensure new updates do not break existing functionality.
   • DFunctional testing to verify the treatment recommendations are clinically accurate.
   Reveal Solution Discussion Select an answer first