ISTQB Certified Tester - Testing with Generative AI Free Sample Questions

Description

Covers generative AI foundations for software testing, developing and refining prompts for test tasks, hallucination, bias, privacy and environmental risks, and adoption roadmaps.

20 free sample questions199 in the full practice test

Try simulator

CT-GenAI Sample Questions

  1. Question 1

    Intermediate

    Introduction to Generative AI for Software Testing · Generative AI Foundations and Key Concepts

    A financial institution is implementing a Generative AI solution to assist in creating test cases for a legacy mainframe system. The system logs are verbose, often exceeding 50,000 words per transaction cycle. The team attempts to paste these logs directly into a standard LLM to identify error patterns, but the model returns incomplete analysis or errors regarding input length. Which fundamental constraint of Large Language Models is the team encountering, and what is the most architecturaly sound resolution?

    Answer and explanation

    Correct answer: C

    The context window limits the amount of text (tokens) the model can process in a single interaction. Mainframe logs often exceed standard limits. Chunking the data or using extended context models is the correct architectural solution.

  2. Question 2

    Intermediate

    Introduction to Generative AI for Software Testing · Generative AI Foundations and Key Concepts

    Which of the following scenarios best illustrates the application of a Multimodal Large Language Model (MLLM) in a software testing context?

    Answer and explanation

    Correct answer: B

    Multimodal models can process multiple types of input simultaneously, such as images (screenshots) and text (requirements). Comparing visual UI elements against textual specs is a prime use case for MLLMs in testing.

  3. Question 3

    Beginner

    Introduction to Generative AI for Software Testing · Generative AI Foundations and Key Concepts

    In the context of Generative AI for software testing, how does 'Tokenization' fundamentally impact the cost and processing of test artifacts?

    Answer and explanation

    Correct answer: C

    Tokenization is the process of converting text into tokens. LLMs process text in tokens, not words. API costs and context window limits are defined in tokens, making it the fundamental unit of consumption.

  4. Question 4

    Intermediate

    Introduction to Generative AI for Software Testing · Generative AI Foundations and Key Concepts

    A test manager is explaining the difference between Discriminative AI and Generative AI to stakeholders. Which comparison accurately reflects their roles in software testing?

    Answer and explanation

    Correct answer: D

    Discriminative models classify data (e.g., pass/fail, spam/not spam), while Generative models create new data (e.g., generating a Python script or a Gherkin scenario) based on training distributions.

  5. Question 5

    Beginner

    Introduction to Generative AI for Software Testing · Leveraging Generative AI in Software Testing: Core Principles

    When using an LLM to generate unit tests, a developer notices the output varies significantly each time the same prompt is submitted. Which parameter should be adjusted to make the output more deterministic for regression testing purposes?

    Answer and explanation

    Correct answer: A

    Temperature controls the randomness of predictions. A lower temperature (near 0) makes the model more deterministic and focused, choosing the most likely next token, which is desirable for code generation consistency.

  6. Question 6

    Intermediate

    Introduction to Generative AI for Software Testing · Generative AI Foundations and Key Concepts

    Which of the following best describes 'Instruction-Tuned' models compared to 'Foundation' models in the context of software testing?

    Answer and explanation

    Correct answer: D

    Foundation models predict the next token based on vast internet data. Instruction-tuned models undergo further training (RLHF or supervised fine-tuning) specifically to learn how to follow user instructions (e.g., 'Act as a QA engineer'), making them superior for testing tasks.

  7. Question 7

    Beginner

    Prompt Engineering for Effective Software Testing · Effective Prompt Development

    A test analyst is crafting a prompt to generate Gherkin scenarios. They explicitly state: "You are a Senior QA Automation Engineer with 10 years of experience in E-commerce." Which component of a structured prompt does this sentence represent?

    Answer and explanation

    Correct answer: C

    This component assigns a specific Role or Persona to the AI. This helps align the tone, vocabulary, and perspective of the output with what is expected from that specific professional profile.

  8. Question 8

    Beginner

    Prompt Engineering for Effective Software Testing · Effective Prompt Development

    You are tasked with using an LLM to generate test cases for a login feature. You initially send a prompt: "Write test cases for login." The output is generic and unusable. You then refine the prompt to include the specific requirements: "The login accepts email and password. Password must be 8 chars. Lockout after 3 attempts." Which prompt engineering principle are you applying to improve the result?

    Answer and explanation

    Correct answer: D

    Adding specific details about the system under test (SUT) provides 'Context'. Without context, the LLM hallucinates generic features. Providing the business rules allows the model to generate relevant test conditions.

  9. Question 9

    Intermediate

    Prompt Engineering for Effective Software Testing · Applying Prompt Engineering Techniques to Software Test Tasks

    A tester wants to generate equivalence partition test data for a purely mathematical function. They provide the LLM with the function logic and then ask for the test data immediately, but the model makes calculation errors. Which prompting technique involves asking the model to "Think step-by-step" to improve logical accuracy?

    Answer and explanation

    Correct answer: C

    Chain-of-Thought prompting encourages the LLM to articulate its reasoning process ('step-by-step') before giving the final answer. This significantly improves performance on logic, math, and reasoning tasks by allowing the model to correct its internal state.

  10. Question 10

    Intermediate

    Prompt Engineering for Effective Software Testing · Effective Prompt Development

    You need to generate test cases that strictly follow a specific JSON format for your test management tool. You provide three complete examples of the input requirement and the corresponding JSON output in your prompt. What technique are you utilizing?

    Answer and explanation

    Correct answer: C

    Few-Shot prompting involves providing specific examples (shots) of the desired input-output pairs within the prompt to guide the model's behavior and formatting. This is highly effective for enforcing structured outputs like JSON.

Register free for 10 more questions

Or unlock all 199 CT-GenAI questions with explanations, timed mode and flashcards.