ISTQB CT-AI: Certified Tester AI Testing (CT-AI) v2.0

Last checked November 10, 2026 against the official ISTQB exam page.

1. Question 1Introduction to Artificial Intelligence

   A startup builds a chatbot that answers questions only about its own product catalogue, and its developers describe it as narrow AI. Which statement best explains this description?

   1. AIt is built to perform specific tasks within a limited domain rather than to match general human intelligence across all tasks
   2. BIt can learn any intellectual task a person can perform and carry its knowledge over to unrelated domains without retraining
   3. CIt has conscious awareness of its own goals, so it can reason about its limitations in every situation it encounters
   4. DIt runs only on local hardware and does not use machine learning, which is why its answers are fixed in advance
   Show answer and explanationHide answer

   Correct answer: A

   Narrow AI, also called weak AI, is designed for a specific task or limited domain. Option B describes general AI, option C attributes a consciousness that the narrow label does not imply, and option D is wrong because narrow AI often uses machine learning and where a system runs does not define its category.

2. Question 2Quality Characteristics for AI-Based Systems

   A navigation app's routing model is retrained every week on new traffic data. The test team must confirm that each new version still meets the quality expectations agreed for the system. Which quality characteristic is most directly concerned with this ongoing change?

   1. AAutonomy, which describes how far the system can act without human intervention
   2. BEvolution, which describes how the system changes over time as it keeps learning from new data
   3. CFlexibility and adaptability, which describes how well the system works in new operating settings
   4. DInterpretability, which describes how well people can understand the reasons behind the model's decisions
   Show answer and explanationHide answer

   Correct answer: B

   Evolution covers how an AI-based system changes as it learns, so each retrained version needs checking against the agreed expectations. Autonomy, flexibility and interpretability are also quality characteristics, but they do not describe change over time caused by retraining.

3. Question 3Machine Learning

   A team trains a model using past emails that people have already labelled as spam or not spam. Which type of machine learning does this describe?

   1. ASupervised learning, because the training data includes a known label for each example
   2. BUnsupervised learning, because the model groups emails by similarity without using any labels
   3. CReinforcement learning, because the model earns rewards for each correct spam decision it makes
   4. DTransfer learning, because the model reuses knowledge gained from a model trained on another task
   Show answer and explanationHide answer

   Correct answer: A

   Learning a mapping from inputs to known outputs using labelled examples defines supervised learning. The other options describe learning without labels, learning from reward signals, and reusing a pretrained model, none of which matches labelled email data.

4. Question 4Testing AI-Based Systems

   A team tests a system that reads handwritten postal codes from envelopes. For an unusual, newly scanned handwriting sample, nobody can state with certainty which code is correct. Which testing challenge does this situation describe?

   1. AThe boundary value problem, where inputs sit at the edges of equivalence partitions
   2. BThe flaky test problem, where the same test passes and fails across repeated runs
   3. CThe test oracle problem, where the correct expected result cannot be determined reliably
   4. DThe regression scope problem, where too many unchanged components must be retested
   Show answer and explanationHide answer

   Correct answer: C

   When no reliable expected result exists for an input, a test oracle cannot be defined, which is the test oracle problem that is common in AI-based systems. The other options name real testing issues, but none of them describes a missing expected result.

5. Question 5Input Data Testing for Machine Learning Systems

   A team reports 98% accuracy for a customer churn model. An audit finds that several customer records appear in both the training data and the test data, with identical feature values. Which concern does this most directly raise?

   1. AUnderfitting, meaning the model is too simple to capture the churn pattern
   2. BData leakage, which inflates the measured performance on the test data
   3. CConcept drift, meaning customer behaviour has changed since the data was collected
   4. DClass imbalance, meaning churners are a small minority of the customer records
   Show answer and explanationHide answer

   Correct answer: B

   When test records also appear in training data, the test result no longer measures performance on unseen data, which is data leakage. Underfitting, concept drift and class imbalance are real issues, but identical records across the split point directly to leakage.