A security operations center uses an AI model to classify network traffic as malicious or benign. The team wants to improve the model's robustness against evasion attacks where attackers modify traffic to avoid detection. Which technique is most effective?

Increase the model's input features to capture more traffic characteristics.

Use ensemble learning by combining multiple models to make predictions.

Deploy the model on a more powerful server to handle larger traffic volumes.

Use a simpler model to reduce the attack surface.

What is the primary purpose of threat modeling for AI systems?

To guarantee the model is 100% accurate

To automatically patch all vulnerabilities in the AI model

To replace the need for continuous monitoring

To systematically identify and prioritize potential adversarial risks and attack vectors

At which stage of the AI development lifecycle should security requirements be first considered to mitigate adversarial risks?

During the design and planning phase

A research institute is building an AI model to analyze scientific papers. The team is concerned about data poisoning, where an attacker could submit fake papers to influence the model's conclusions. They want to ensure the integrity of the training data without significantly increasing the data collection effort. Which control would be most effective?

Use a larger dataset to dilute the impact of any poisoned data.

Encrypt the data to prevent unauthorized access.

Implement a peer-review process for all data sources before they are included in the training set.

Use a more complex model to better distinguish between real and fake papers.

CompTIASecAI+

Domain 2Objective 3

Mitigate Adversarial Risks CY0-001 Practice Questions (Page 5)

Part of the Securing AI systems domain, which accounts for 40% of the CY0-001 exam. CompTIA does not publish an official question count, but from its 60-minute exam (~25–40 total, ~10–16 in this domain), expect 3–5 from this objective — we provide 34 practice questions to prepare you well beyond it. (estimate)

34questions here

7free pages

8concepts

40%of the exam

Questions 21–25

21application · medium
A security operations center uses an AI model to classify network traffic as malicious or benign. The team wants to improve the model's robustness against evasion attacks where attackers modify traffic to avoid detection. Which technique is most effective?
Select an answer first
22foundation · easy
What is the primary purpose of threat modeling for AI systems?
Select an answer first
23foundation · easy
At which stage of the AI development lifecycle should security requirements be first considered to mitigate adversarial risks?
Select an answer first
24expert · hard
A research institute is building an AI model to analyze scientific papers. The team is concerned about data poisoning, where an attacker could submit fake papers to influence the model's conclusions. They want to ensure the integrity of the training data without significantly increasing the data collection effort. Which control would be most effective?
Select an answer first
25foundation · easy
Which monitoring technique involves tracking the distribution of input data to detect when it deviates significantly from the training distribution?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to CY0-001

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by CompTIA. “CY0-001” is a trademark of its owner, used for identification only.

Mitigate Adversarial Risks CY0-001 Practice Questions (Page 5)

Questions 21–25

A security operations center uses an AI model to classify network traffic as malicious or benign. The team wants to improve the model's robustness against evasion attacks where attackers modify traffic to avoid detection. Which technique is most effective?

What is the primary purpose of threat modeling for AI systems?

At which stage of the AI development lifecycle should security requirements be first considered to mitigate adversarial risks?

Which monitoring technique involves tracking the distribution of input data to detect when it deviates significantly from the training distribution?

Keep studying — Securing AI systems