Page 1 of 63
Testing with Generative AI (CT-Genial)
Exam Questions And Correct Answers
(Verified Answers) Plus Rationales 2026
Q&A | Instant Download Pdf.
Question 1
What is the primary characteristic that distinguishes generative
AI from discriminative AI?
A) Generative AI only classifies existing data
B) Generative AI creates new content that resembles the training
distribution
C) Generative AI requires structured data only
D) Generative AI cannot be used for testing purposes
Answer: B
,Page 2 of 63
Rationale: Generative AI models learn the underlying
probability distribution of training data and can generate new
samples. Discriminative AI focuses on distinguishing between
categories (classification). The ability to create novel, realistic
outputs is the key differentiator .
Question 2
Which of the following is a typical application of generative AI
in software testing?
A) Static code analysis without ML
B) Generating synthetic test data
C) Manual test case design
D) Running performance tests
Answer: B
,Page 3 of 63
Rationale: Generative AI excels at producing realistic synthetic
data (text, images, tabular) to augment test datasets, simulate
edge cases, and preserve privacy. The other options do not
leverage generative capabilities .
Question 3
In the context of testing, a "test oracle" for a generative AI
system is:
A) The training dataset used to build the model
B) A mechanism to determine whether generated output is correct
C) The prompt template used by testers
D) The hardware acceleration for inference
Answer: B
Rationale: The test oracle is a principle or mechanism that
decides if a test passes or fails. For generative AI, oracles are
, Page 4 of 63
often non-trivial because outputs are probabilistic and lack a
single expected value .
Question 4
What does "hallucination" mean in generative AI testing?
A) The model refuses to answer a prompt
B) The model generates factually incorrect or nonsensical content
confidently
C) The model runs out of memory during inference
D) The model produces perfect outputs every time
Answer: B
Rationale: Hallucination is a well-known failure mode where a
generative model invents information that is not true or not
grounded in its training data. Testing must detect and mitigate
such outputs .
Testing with Generative AI (CT-Genial)
Exam Questions And Correct Answers
(Verified Answers) Plus Rationales 2026
Q&A | Instant Download Pdf.
Question 1
What is the primary characteristic that distinguishes generative
AI from discriminative AI?
A) Generative AI only classifies existing data
B) Generative AI creates new content that resembles the training
distribution
C) Generative AI requires structured data only
D) Generative AI cannot be used for testing purposes
Answer: B
,Page 2 of 63
Rationale: Generative AI models learn the underlying
probability distribution of training data and can generate new
samples. Discriminative AI focuses on distinguishing between
categories (classification). The ability to create novel, realistic
outputs is the key differentiator .
Question 2
Which of the following is a typical application of generative AI
in software testing?
A) Static code analysis without ML
B) Generating synthetic test data
C) Manual test case design
D) Running performance tests
Answer: B
,Page 3 of 63
Rationale: Generative AI excels at producing realistic synthetic
data (text, images, tabular) to augment test datasets, simulate
edge cases, and preserve privacy. The other options do not
leverage generative capabilities .
Question 3
In the context of testing, a "test oracle" for a generative AI
system is:
A) The training dataset used to build the model
B) A mechanism to determine whether generated output is correct
C) The prompt template used by testers
D) The hardware acceleration for inference
Answer: B
Rationale: The test oracle is a principle or mechanism that
decides if a test passes or fails. For generative AI, oracles are
, Page 4 of 63
often non-trivial because outputs are probabilistic and lack a
single expected value .
Question 4
What does "hallucination" mean in generative AI testing?
A) The model refuses to answer a prompt
B) The model generates factually incorrect or nonsensical content
confidently
C) The model runs out of memory during inference
D) The model produces perfect outputs every time
Answer: B
Rationale: Hallucination is a well-known failure mode where a
generative model invents information that is not true or not
grounded in its training data. Testing must detect and mitigate
such outputs .
