Free Certified Tester Testing with Generative AI Practice Questions

Sample Practice Questions

Try these 10 sample questions from the Certified Tester Testing with Generative AI practice bank. Each includes the answer and explanation. Continue with the interactive practice above to review the full question set.

1. 1Which approach represents knowledge explicitly as symbols and logical rules?

   1. A. Classical machine learning
   2. B. Symbolic artificial intelligence
   3. C. Deep learning
   4. D. Generative artificial intelligence

   Answer: B

   Explanation: The syllabus states, "Symbolic AI uses a rule-based system to mimic human decision-making." Symbolic AI encodes explicit knowledge and rules; the other approaches learn patterns from data.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 1.1.1, p.14

2. 2What purpose does the role component serve in a structured testing prompt?

   1. A. It supplies the source records to be analyzed.
   2. B. It specifies the required response schema.
   3. C. It restricts which input fields may be used.
   4. D. It establishes the perspective the model should adopt.

   Answer: D

   Explanation: The syllabus states, "Role: The role defines the perspective or persona that the GenAI model should take when generating a response." Role establishes a persona or professional perspective; data, formatting, and restrictions belong to other prompt components.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 2.1.1, p.21

3. 3An LLM-generated test script calls a plausible-sounding API method. The API reference and implementation confirm that the method does not exist. Which output defect does this illustrate?

   1. A. Training-data bias that favors one group of test users
   2. B. A reasoning error in the ordering of documented API steps
   3. C. Nondeterministic wording between otherwise valid test scripts
   4. D. A hallucination that invents an operation of the tested interface

   Answer: D

   Explanation: The syllabus explains that "hallucinations can manifest as LLMs creating fictitious or irrelevant test cases, generating incorrect or non-functioning test scripts, or suggesting test cases that verify nonexistent acceptance criteria." Inventing an interface operation is fictitious content. The evidence is not unequal user-group coverage, incorrect sequencing of real operations, or merely different wording of valid scripts.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 3.1.1, p.35

4. 4In the syllabus's typical LLM-powered test architecture, which component manages authentication, data retrieval, and prompt preparation?

   1. A. The front-end user interface
   2. B. The language model's embedding vectors
   3. C. The application back-end
   4. D. The model's original training corpus

   Answer: C

   Explanation: The syllabus states, "The back-end processes user input and manages critical functions such as authentication, data retrieval, prompt preparation, and interaction with the LLM." The back-end coordinates these functions. The front-end captures interaction, while embeddings and training data are representations or resources rather than the coordinating application component.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 4.1.1, p.44

5. 5A shadow AI tool has unclear licensing terms, and a tester plans to submit copyrighted requirements. Which conclusion about this use is justified?

   1. A. Technical access to the tool establishes permission to process copyrighted inputs.
   2. B. Keeping generated testware inside the company removes copyright-authorization concerns.
   3. C. Processing copyrighted material without authorization can lead to intellectual property disputes.
   4. D. A paid user account establishes ownership rights to every generated testing artifact.

   Answer: C

   Explanation: The syllabus states, "The use of AI tools with unclear licensing agreements can expose LLM users to intellectual property disputes, especially if copyrighted data is processed without proper authorization." Unclear authorization creates dispute risk. Technical access, internal use, and payment do not themselves establish copyright permission or ownership rights.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 5.1.1, p.50

6. 6How does classical machine learning typically prepare information for model training?

   1. A. It replaces training data with a fixed set of logical rules.
   2. B. It generates new examples instead of selecting input features.
   3. C. It automatically learns all features through deep neural layers.
   4. D. It uses data preparation and selection of relevant features.

   Answer: D

   Explanation: The syllabus states, "Classical machine learning is a data-driven approach that requires data preparation, feature selection and model training". Selecting relevant features distinguishes this preparation from symbolic rules and automatic feature learning in deep learning.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 1.1.1, p.14

7. 7A prompt explains that the test object is a ticketing service used during high-demand sales. Which component supplies this background?

   1. A. Context
   2. B. Role
   3. C. Output format
   4. D. Instruction

   Answer: A

   Explanation: The syllabus states, "Context: Context provides the background information that the GenAI model needs to determine the test conditions." Background about the test object and its operating situation is context, rather than a persona, response layout, or directive.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 2.1.1, p.21

8. 8The supplied rule grants access only when an account is active AND its subscription is paid. An LLM quotes both facts correctly but concludes that an active, unpaid account qualifies because one condition is true. Which specific error explains this conclusion?

   1. A. A reasoning error that treats a conjunction as an alternative
   2. B. A training-data bias favoring one demographic group
   3. C. A disclosure of confidential information from training data
   4. D. A tokenization error caused by an unfamiliar account name

   Answer: A

   Explanation: The syllabus states, "Reasoning errors occur when LLMs misinterpret logical structures, such as cause-and-effect relationships, conditional logic, or step-by-step problem-solving processes, leading to incorrect conclusions." The model misapplies AND despite retaining the supplied facts. The evidence concerns conditional logic, rather than demographic preference, disclosure, or token splitting.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 3.1.1, p.35

9. 9What role does a vector database serve in LLM-powered test infrastructure?

   1. A. Retrieving related content by semantic similarity of embeddings
   2. B. Looking up documents by exact equality between query words and keywords
   3. C. Storing learned model weights for use in subsequent training updates
   4. D. Returning cached complete answers by matching stored request identifiers

   Answer: A

   Explanation: The syllabus lists "Vector databases (for semantic retrieval of related content using embeddings; see section 4.1.2)." The characteristic role is embedding-based semantic retrieval. Exact keyword lookup confuses semantic retrieval with lexical matching; learned weights belong to the model; request-keyed complete answers describe a response cache.

   Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 4.1.1, p.44

10. 10Which organizational guidelines support responsible GenAI adoption in the test process?

    1. A. Rules for sensitive data, disclosure of AI use, and review gates for testware
    2. B. Rules for buying licenses, with output acceptance based on model confidence alone
    3. C. Rules for standardizing report layout, with sensitive-data decisions left to each user
    4. D. Rules for counting prompts, with generated testware accepted whenever it is fluent

    Answer: A

    Explanation: The syllabus states that "organizations should establish process guidelines for the use of GenAI, including rules for the use of sensitive data, transparency obligations (e.g., what was generated using GenAI), and quality gates with review of generated testware." Adoption guidance includes data handling, transparency, and quality review. Licensing, formatting, usage counts, fluency, and confidence alone do not provide those controls.

    Reference: ISTQB Certified Tester Testing with Generative AI Syllabus v1.1 — Section 5.1.2, p.50