AI Capabilities, Data Dependencies, and Failure Modes CAIPM Practice Questions (Page 6)

Questions 26–30

1. 26expert · hard

   A weather forecasting company uses an AI model to predict rainfall. The model was trained on data from the past decade and performs well in most regions. However, in a region that recently experienced a change in climate patterns, the model's predictions have become unreliable. The company has limited budget for retraining. What is the most cost-effective first step?

   • AIncrease the model's complexity to capture the new climate patterns.
   • BImplement monitoring to detect data drift and quantify the performance drop in that region.
   • CReplace the model with a simpler statistical model for that region.
   • DImmediately retrain the model on the most recent data from that region.
   Reveal Solution Discussion Select an answer first
2. 27expert · hard

   A bank uses an AI system to approve credit card transactions in real time. The system was trained on historical transaction data. Recently, fraudsters have changed their patterns, and the system is now blocking many legitimate transactions while missing some fraudulent ones. The bank wants to reduce both false positives and false negatives, but the fraud patterns are evolving rapidly. What is the most effective strategy?

   • AIncrease the decision threshold to block more transactions, reducing fraud but increasing false positives.
   • BImplement an ensemble of models that are retrained on a rolling basis with recent fraud data, and use online learning to adapt to new patterns.
   • CSwitch to a rule-based system with manually updated rules for known fraud patterns.
   • DUse a single deep learning model trained on all historical data and do not retrain frequently.
   Reveal Solution Discussion Select an answer first
4. 28application · easy

   A logistics company wants to use AI to automatically read addresses from handwritten package labels and route packages to the correct destination. Which AI capability is most directly required for this task?

   • ALearning to adapt to new handwriting styles over time.
   • BPerception, specifically optical character recognition and handwriting recognition.
   • CNatural language processing to understand the meaning of the address text.
   • DReasoning to infer the correct route from the address.
   Reveal Solution Discussion Select an answer first
5. 29foundation · easy

   An AI model for predicting customer churn was trained on data from 2020. In 2024, the model's accuracy drops significantly because customer behavior patterns have changed. Which failure mode is this?

   • AData drift
   • BModel degradation
   • CBias
   • DOverfitting
   Reveal Solution Discussion Select an answer first
6. 30foundation · easy

   Which AI capability is most directly applied when a system analyzes customer reviews to determine whether the sentiment expressed is positive, negative, or neutral?

   • APerception
   • BLearning
   • CNatural language processing
   • DReasoning
   Reveal Solution Discussion Select an answer first