View Full ISTQB CT-GenAI Exam Dumps and Practice Test Dumps.

 

Question 221

A tester asks a GenAI tool to generate API test cases. The prompt includes the endpoint, HTTP methods, authentication rules, required fields, and expected response codes. What is the primary benefit of providing this context?

1. It guarantees that every generated test will pass
2. It removes the need for human review
3. It gives the model information needed to generate more relevant tests
4. It prevents all possible hallucinations

Correct Answer: 3

Explanation

Providing relevant context helps a GenAI tool understand the specific system behavior that the tester wants to verify. Endpoint information, supported methods, authentication requirements, fields, and expected response codes give the model useful boundaries for generating test scenarios. Without this information, the model may make assumptions about the API and produce irrelevant or inaccurate tests. Context does not guarantee correctness, eliminate hallucinations, or remove the need for human review. The generated cases should still be checked against authoritative API specifications and actual application behavior. Therefore, the main benefit is improved relevance and alignment of the generated tests with the intended API requirements.

Question 222

A GenAI tool generates several test cases that appear almost identical. Which approach would best help the tester identify unnecessary duplication?

1. Compare the generated cases against existing test objectives and coverage
2. Increase the temperature without reviewing the output
3. Remove all expected results from the test cases
4. Ask the tool to generate unlimited additional cases

Correct Answer: 1

Explanation

Duplicate or highly similar test cases can increase maintenance effort without providing meaningful additional coverage. Comparing generated cases against existing test objectives, scenarios, and coverage helps identify whether each case contributes a distinct testing purpose. Simply generating more cases can increase duplication, while removing expected results reduces test quality. Changing model parameters may alter output variety but does not establish whether the resulting cases are necessary. The tester should assess equivalence in objectives, input conditions, expected behavior, and coverage. This review also helps maintain a manageable test suite and ensures that GenAI-generated content adds useful testing value rather than simply increasing the number of test cases.

Question 223

Which prompt is most appropriate when asking GenAI to generate negative tests for a user registration feature?

1. Generate some registration tests.
2. Create tests without considering requirements.
3. Generate only successful registration scenarios.
4. Generate negative tests covering invalid email formats, missing mandatory fields, weak passwords, and duplicate accounts.

Correct Answer: 4

Explanation

A strong prompt identifies the specific negative-testing areas that should be addressed. Invalid email formats, missing mandatory fields, weak passwords, and duplicate accounts represent concrete categories of invalid or undesirable input. This gives the GenAI tool clear boundaries and increases the likelihood that the generated tests will be relevant to the intended test objective. A vague request such as “generate some tests” provides insufficient guidance, while requesting only successful scenarios excludes negative testing entirely. The resulting cases should still be reviewed against the actual registration requirements because GenAI can omit important conditions or introduce unsupported assumptions.

Question 224

A tester receives generated test cases containing a requirement ID that does not exist in the approved requirements repository. What should the tester do first?

1. Add the invented requirement ID to the test management system
2. Verify the requirement ID against the authoritative requirements source
3. Assume the requirement was recently created
4. Execute the test case without checking the requirement

Correct Answer: 2

Explanation

An invented or incorrect requirement ID is an example of potentially hallucinated information. The tester should verify the identifier against an authoritative requirements repository before using it for traceability or test execution. Adding an unsupported identifier could create false traceability and make future reporting inaccurate. Assuming that the requirement exists or executing the test without verification does not address the underlying issue. GenAI-generated references should be treated as output requiring validation rather than automatically trusted facts. Once the correct requirement information has been confirmed, the tester can revise the test case and establish legitimate traceability to the approved specification.

Question 225

A team wants to use GenAI to create synthetic customer records for testing while avoiding exposure of real customer information. What is a major advantage of this approach?

1. It can reduce the need to use sensitive production data
2. It guarantees that the synthetic data perfectly represents production
3. It eliminates all data-quality testing
4. It makes privacy controls unnecessary

Correct Answer: 1

Explanation

Synthetic data can provide realistic-looking test records without directly exposing real customer information. This can reduce dependence on production data containing personally identifiable or otherwise sensitive information. However, synthetic data is not automatically equivalent to production data and must be evaluated for suitability, diversity, validity, and coverage of relevant business conditions. Privacy controls and organizational policies still apply. Data generated by GenAI should also be checked for accidental reproduction of sensitive information or unrealistic patterns. Therefore, the major advantage is that synthetic data can reduce the need to use sensitive production records while supporting controlled testing activities.

Question 226

Which characteristic is most important when evaluating whether a GenAI-generated test case is suitable for inclusion in an approved test suite?

1. Whether the test case uses complex technical terminology
2. Whether the test case is longer than existing tests
3. Whether the test case aligns with an approved test objective or requirement
4. Whether the test case was generated quickly

Correct Answer: 3

Explanation

A generated test case should provide meaningful testing value and align with an approved requirement, risk, or test objective. Technical complexity, length, and generation speed do not demonstrate that the case is appropriate for inclusion. The tester should assess whether the scenario is relevant, accurate, sufficiently detailed, and consistent with the system specification. Expected results and test data should also be reviewed for correctness. Alignment with approved objectives helps prevent irrelevant or unsupported tests from entering the formal test suite. GenAI can accelerate test creation, but the generated output still requires professional evaluation before it becomes part of controlled testing activities.

Question 227

A tester uses a GenAI assistant to transform existing manual test cases into automation scripts. What is an important validation activity after code generation?

1. Confirm that the generated code preserves the original test intent and expected behavior
2. Assume the generated script is correct because it compiles
3. Remove all assertions to simplify maintenance
4. Replace the approved test case with the generated code

Correct Answer: 1

Explanation

Generated automation code must preserve the intent of the original manual test. A script may compile successfully while still using incorrect locators, missing important steps, checking the wrong expected result, or failing to represent required business rules. Therefore, the tester should compare the generated script with the source test case and verify its inputs, actions, assertions, data, and expected outcomes. Successful compilation is not sufficient evidence of correctness. The generated code should also be reviewed for maintainability, security, and compatibility with the automation framework. Human validation remains important when GenAI is used to accelerate automation development.

Question 228

A GenAI system is asked to summarize test execution results. Which information should the tester verify before publishing the generated summary?

1. Only the grammar and spelling
2. The factual accuracy of pass, fail, blocked, and other execution results
3. Only the number of words in the summary
4. The model’s preferred writing style

Correct Answer: 2

Explanation

A test execution summary must accurately represent the actual execution evidence. Pass, fail, blocked, skipped, and other statuses should be checked against authoritative test-management or execution records. A grammatically perfect summary can still be misleading if the underlying figures or statuses are incorrect. The tester should also verify important defect references, environment information, scope, and notable limitations where applicable. GenAI is useful for organizing and communicating information, but it should not be treated as the authoritative source of execution results. Verification against actual records helps prevent hallucinated metrics, incorrect conclusions, and misleading reports from reaching stakeholders.

Question 229

A tester notices that a GenAI-generated answer changes between two identical requests. Which property of GenAI output does this illustrate?

1. Deterministic execution
2. Guaranteed reproducibility
3. Static rule processing
4. Probabilistic or non-deterministic generation

Correct Answer: 4

Explanation

Generative AI systems can produce different outputs for identical or very similar prompts because generation is generally probabilistic. Depending on the model and configuration, variations may occur in wording, examples, ordering, or even substantive content. This behavior means that testers should not assume that repeating a prompt will always produce exactly the same result. For important testing activities, teams may need to preserve prompts, model versions, relevant configuration, and generated outputs to support reproducibility and auditability. The presence of output variation does not automatically indicate a defect, but it does require appropriate validation when consistency is important to the testing task.

Question 230

A tester wants GenAI to generate test scenarios specifically for a payment service used by administrators. Which prompt element would most improve relevance?

1. Asking for the longest possible response
2. Removing all information about users
3. Specifying the administrator role and payment-related business context
4. Asking the model to make assumptions about the application

Correct Answer: 3

Explanation

Providing the intended user role and business context helps the model generate scenarios that are relevant to the actual system behavior. Administrator permissions may differ significantly from those of ordinary users, especially for payment operations. Including payment rules, relevant roles, constraints, and expected behavior can further improve the usefulness of the generated scenarios. Asking the model to make assumptions can introduce unsupported behavior, while removing context reduces relevance. Longer responses do not necessarily provide better tests. The generated scenarios should still be compared with approved authorization requirements and payment specifications before being accepted into the test process.

Question 231

Which situation is an example of prompt injection affecting a GenAI-assisted testing workflow?

1. A tester provides valid acceptance criteria
2. A malicious input attempts to instruct the model to ignore its testing instructions and reveal restricted information
3. A tester asks for boundary-value test cases
4. A tester requests a structured JSON response

Correct Answer: 2

Explanation

Prompt injection occurs when untrusted content attempts to influence the behavior of a GenAI system by inserting instructions that conflict with the intended task. In a testing workflow, malicious or unexpected content could attempt to make the model ignore its original instructions, disclose confidential information, or perform an unauthorized action. Acceptance criteria, boundary-value requests, and structured output instructions are normal prompt elements when they are part of the intended workflow. Testers should therefore treat external or untrusted content as potentially unsafe and establish controls around what information the model can access and what actions it can perform.

Question 232

A tester asks GenAI to review requirements for testability. Which finding would be most useful for the tester to investigate?

1. A requirement uses an ambiguous term without defining measurable behavior
2. The requirement contains a short sentence
3. The requirement uses a familiar business word
4. The requirement has been stored electronically

Correct Answer: 1

Explanation

Ambiguous requirements can make it difficult to determine what should be tested and what constitutes a successful result. A requirement containing an undefined term such as “quickly,” “securely,” or “user-friendly” may lack measurable acceptance criteria. GenAI can help identify potentially ambiguous language and suggest questions for clarification, but its findings should be reviewed by qualified stakeholders. Sentence length, storage format, or the use of familiar terminology does not automatically indicate a testability problem. The tester should use the generated observations as review assistance and confirm the actual interpretation with the appropriate product, business, or requirements owner.

Question 233

A company policy prohibits sending proprietary source code to external GenAI services. What should a tester do when an external tool requests the code to generate unit tests?

1. Upload the source code because unit tests are low risk
2. Remove only comments and upload the remaining code
3. Follow the approved policy and use an authorized alternative
4. Ignore the policy if the generated tests are useful

Correct Answer: 3

Explanation

Organizational policies concerning proprietary source code and external GenAI services must be followed. Uploading restricted source code can create confidentiality, intellectual-property, security, or contractual risks. Removing comments may not remove sensitive implementation details, so partial redaction is not automatically sufficient. A tester should instead use an approved internal model, an authorized service, sanitized examples, or another permitted approach. If uncertainty exists, the tester should consult the responsible security, legal, or governance function. GenAI productivity benefits do not override organizational restrictions. Appropriate controls ensure that testing assistance is obtained without exposing protected source code to unauthorized systems.

Question 234

Which approach provides the strongest basis for assessing whether GenAI-generated test cases improve testing effectiveness?

1. Count only how many test cases were generated
2. Measure relevant quality and coverage outcomes against defined criteria
3. Assume effectiveness because testers saved time
4. Use the model’s own statement that its tests are comprehensive

Correct Answer: 2

Explanation

Effectiveness should be assessed using predefined and relevant criteria rather than simply counting generated cases. Depending on the context, these criteria may include requirement coverage, defect detection, scenario diversity, review effort, execution quality, or maintenance impact. Time savings can be useful, but they do not by themselves demonstrate improved testing effectiveness. Similarly, the GenAI tool cannot independently establish that its output is comprehensive. A controlled evaluation comparing appropriate outcomes can provide stronger evidence of value. The criteria should reflect the organization’s testing objectives and should account for both benefits and risks introduced by GenAI-assisted test generation.

Question 235

A tester asks GenAI to generate test data for an age field that accepts values from 18 through 65 inclusive. Which data set would best support boundary testing?

1. 18, 19, 64, 65
2. 1, 10, 30, 90
3. 20, 30, 40, 50
4. 25, 35, 45, 55

Correct Answer: 1

Explanation

For an age field with an inclusive valid range from 18 to 65, boundary testing should focus on values at and around the boundaries. Values such as 18 and 65 represent valid boundary conditions, while 19 and 64 are immediately inside the valid range. Depending on the test design technique, values just outside the range, such as 17 and 66, would also be useful for negative boundary testing. The key point is that generated data should be derived from the approved requirement rather than arbitrary values. The tester should verify that the generated test data reflects the specified boundary conditions.

Question 236

A tester uses GenAI to suggest possible causes for a failed test. What is the appropriate way to treat these suggestions?

1. Treat every suggested cause as the confirmed root cause
2. Use the suggestions as hypotheses that require investigation
3. Record all suggestions as confirmed defects
4. Ignore the actual failure evidence

Correct Answer: 2

Explanation

GenAI can help testers brainstorm possible explanations for a failure, but generated explanations are hypotheses rather than confirmed root causes. The tester should investigate logs, application behavior, configuration, recent changes, environment conditions, and other available evidence before determining the actual cause. Treating every generated suggestion as fact could result in incorrect defect reports and wasted investigation effort. GenAI is useful for expanding the range of possible explanations, especially during exploratory analysis, but authoritative evidence must determine the final conclusion. This distinction between generated hypotheses and verified findings is important for maintaining accurate defect analysis and reporting.

Question 237

A tester wants GenAI to produce a defect report that can be imported into a structured system. Which prompt instruction is most useful?

1. “Write something professional.”
2. “Describe the problem in any format.”
3. “Make the report as detailed as possible.”
4. “Return valid JSON containing summary, steps, expected result, actual result, severity, and environment.”

Correct Answer: 4

Explanation

Structured output requirements can make GenAI responses easier to review, process, and integrate with other tools. Specifying valid JSON and defining the required fields gives the model clear formatting constraints. A vague request about professionalism or detail does not ensure consistent structure. Even when valid JSON is produced, the tester must still verify the factual content of each field. Severity, expected results, actual results, and environment details should be based on available evidence rather than model assumptions. Structured prompting therefore supports consistency and automation, while human validation remains necessary before a generated defect report is submitted.

Question 238

A tester discovers that GenAI-generated tests cover common user journeys but omit an uncommon administrative workflow. What should the tester consider?

1. Whether the prompt and source requirements provided enough information about the administrative workflow
2. Whether the model should automatically be trusted to cover all workflows
3. Whether uncommon workflows should always be excluded
4. Whether generating more copies of the common scenarios will solve the problem

Correct Answer: 1

Explanation

GenAI output is influenced strongly by the context provided in the prompt and source material. If an uncommon administrative workflow was not clearly represented, the model may prioritize more common scenarios. The tester should review the requirements, business rules, roles, and prompt context to determine whether the missing workflow was adequately specified. Additional targeted instructions or examples can then be used to improve coverage. The model should not be assumed to provide complete coverage automatically. Uncommon workflows may still be important, particularly when they involve privileged access, financial operations, security controls, or other high-risk functionality.

Question 239

Which practice best supports accountability when GenAI is used to create test artifacts?

1. Allow the model to approve its own generated tests
2. Remove the tester’s name from the final artifacts
3. Keep human review and clearly defined responsibility for approving the artifacts
4. Accept all generated artifacts without modification

Correct Answer: 3

Explanation

Human accountability remains important when GenAI contributes to testing activities. A defined person or role should remain responsible for reviewing and approving generated test cases, scripts, summaries, or other artifacts. The model itself cannot assume organizational accountability for the correctness or suitability of its output. Maintaining review responsibilities also helps identify hallucinations, unsupported assumptions, security problems, and requirement mismatches before artifacts are formally used. Organizations may additionally preserve relevant prompts, model information, review records, and approval history depending on their governance requirements. GenAI can assist with creation, but responsibility for the final testing decision remains with authorized human personnel.

Question 240

A tester wants to reduce hallucinations when asking GenAI to create tests from a requirements document. Which prompting approach is most appropriate?

1. Ask the model to invent reasonable requirements where information is missing
2. Instruct the model to use only the supplied requirements and explicitly identify missing information
3. Remove all requirements from the prompt
4. Ask for the maximum possible number of tests without constraints

Correct Answer: 2

Explanation

Restricting the model to supplied requirements and requiring it to identify missing information can reduce unsupported assumptions. This approach gives the model a clear source of truth and creates a mechanism for exposing gaps instead of silently filling them with invented details. Asking the model to invent requirements increases the risk of hallucinated behavior, while removing requirements eliminates important context. Generating unlimited tests without constraints can also increase irrelevant or unsupported output. Even with a carefully constrained prompt, the resulting tests should be validated against the authoritative requirements because prompt controls reduce risk but cannot guarantee that every generated statement or test case is correct.