Flexibility and Adaptability CT-AI Practice Questions (Page 4)

Questions 16–20

1. 16expert · hard

   A navigation app for drivers uses real-time traffic data to suggest routes. The app must be flexible enough to handle sudden road closures and accidents. However, the team is concerned that too much flexibility might cause the app to frequently change routes, confusing users. Which trade-off is the team considering?

   • AFlexibility vs. robustness, because the app must handle errors in traffic data.
   • BAdaptability vs. reliability, because the app must learn from past traffic patterns.
   • CAdaptability vs. explainability, because the app must explain route changes.
   • DFlexibility vs. reliability, because frequent route changes may reduce user trust in the app's consistency.
   Reveal Solution Discussion Select an answer first
2. 17expert · hard

   A security surveillance system uses AI to detect suspicious behavior. The system is deployed in a public area where lighting and crowd density vary. The team wants the system to be flexible enough to handle these variations. However, they also need the system to be robust against adversarial attacks. Which trade-off is the team facing?

   • AFlexibility vs. robustness, because making the system more flexible to input variations may increase its vulnerability to adversarial inputs.
   • BAdaptability vs. robustness, because the system must update its models.
   • CAdaptability vs. explainability, because the system must learn from new data.
   • DFlexibility vs. reliability, because the system must be consistent.
   Reveal Solution Discussion Select an answer first
4. 18application · medium

   A spam filter for an email service uses a machine learning model. Spammers constantly change their tactics, so the filter must update its model to continue blocking new spam patterns. Which approach best supports the filter's adaptability?

   • ATest the model with a static set of spam examples.
   • BFreeze the model and rely on user reports to manually block spam.
   • CImplement an online learning mechanism that updates the model with new spam and ham examples.
   • DIncrease the model's sensitivity to catch more spam.
   Reveal Solution Discussion Select an answer first
5. 19application · medium

   A voice assistant is used in a call center. Customers speak with various accents and dialects. The assistant's speech recognition system was trained on a standard accent. The team wants to test whether the assistant can understand different accents without retraining. Which testing approach is most suitable?

   • ACreate a test set with audio samples from speakers with diverse accents and dialects.
   • BEvaluate the assistant's response time under different network conditions.
   • CTest the assistant only with the standard accent to ensure baseline performance.
   • DRetrain the model with accent-specific data before testing.
   Reveal Solution Discussion Select an answer first
6. 20application · medium

   A retail company's AI chatbot answers customer queries about products and orders. During the holiday season, the chatbot must handle a sudden surge in queries about gift wrapping and return policies, topics it rarely encountered before. The chatbot is rule-based and does not update its knowledge base. Which quality characteristic is most directly challenged in this scenario?

   • AAdaptability, because the chatbot must learn from the new queries to improve future responses.
   • BExplainability, because the chatbot must justify its responses to customers.
   • CFlexibility, because the chatbot must handle new input variations without changing its underlying rules.
   • DRobustness, because the chatbot must maintain performance despite noisy or malformed inputs.
   Reveal Solution Discussion Select an answer first