A security team is designing defenses for a model that classifies emails as spam or not. They want to protect against both adversarial inputs and data poisoning. Which combination of controls is most effective?

Input validation and output filtering

Model retraining and continuous monitoring

Adversarial training and data provenance checks

What is the purpose of adversarial training as a mitigation strategy?

To make the model more robust to adversarial inputs by including them in the training data

To encrypt the model's parameters

To simplify the model's architecture

What is the primary objective of a data poisoning attack?

To intercept data during model inference

To overload the training infrastructure

To corrupt the training data so the model learns incorrect patterns

To steal the model's parameters

What is the primary goal of an adversarial attack on an AI model?

To cause the model to make incorrect predictions

To steal the model's training data

To replace the model with a malicious version

To overload the model with requests

ISACAAdvanced in AI Audit

Domain 2Objective 6

Threats and Vulnerabilities Specific to AI AAIA Practice Questions (Page 4)

Part of the AI Operations domain, which accounts for 46% of the AAIA exam.

31questions here

7free pages

8concepts

46%of the exam

Questions 16–20

16expert · hard
A security team is designing defenses for a model that classifies emails as spam or not. They want to protect against both adversarial inputs and data poisoning. Which combination of controls is most effective?
Select an answer first
17application · medium
A retail company deploys a demand forecasting model. Over six months, the model's accuracy declines because customer behavior shifts due to a new competitor. The company does not have a process to detect this decline and only notices after significant financial loss. Which AI-specific deployment vulnerability is this?
Select an answer first
18foundation · easy
What is the purpose of adversarial training as a mitigation strategy?
Select an answer first
19foundation · easy
What is the primary objective of a data poisoning attack?
Select an answer first
20foundation · easy
What is the primary goal of an adversarial attack on an AI model?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AAIA

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 ISACA. “AAIA” is a trademark of its owner, used for identification only.