In the context of AI fundamentals, what is the primary role of a neural network in a machine learning system?

To provide a graphical interface for security analysts to interact with AI models

To store labeled training data for later retrieval

To compute a mapping from inputs to outputs by learning patterns from data

To replace the need for feature engineering by hard-coding rules

A red team is testing an AI-based intrusion detection system (IDS) that uses a neural network to classify network flows. The team wants to evaluate whether an attacker could evade detection without causing noticeable changes to the traffic. Which approach best simulates a realistic evasion attempt while keeping the test within scope?

Replace the neural network with a simpler rule-based classifier to see if it is easier to bypass.

Craft adversarial examples by adding small perturbations to malicious flows that the model misclassifies as benign.

Flood the network with high-volume traffic to overwhelm the IDS and cause it to drop packets.

Poison the training data by injecting benign-labeled malicious samples before the IDS is deployed.

Which method is commonly used in adversary emulation to test AI-driven security systems?

Running a red team exercise that follows a known attack framework such as MITRE ATT&CK

Deploying a honeypot with no monitoring or logging

Performing a vulnerability scan with default credentials

Reviewing system logs after an attack has already occurred

A security team deploys an AI-based email gateway that filters phishing messages. The model was trained on a public dataset of emails from 2020–2022. After deployment, the gateway's false-negative rate for phishing increases sharply over two months. The team confirms the model's code and infrastructure are unchanged. Which action best addresses the most likely root cause?

Increase the frequency of model inference to process emails faster

Roll back to the previous version of the model

Add more CPU resources to the inference server

Retrain the model on recent email samples that reflect the current threat landscape

GIAC AI Security Automation Engineer

Domain 3Objective 1

Artificial Intelligence Fundamentals GASAE Practice Questions (Page 5)

Part of the AI and Adversary Emulation domain, which makes up ~19% of our current practice bank. GIAC (SANS) does not publish an official question count, but from its 180-minute exam (~70–120 total, ~13–23 in this domain), expect 7–12 from this objective — we provide 36 practice questions to prepare you well beyond it. (estimate)

36questions here

8free pages

5concepts

Questions 21–25

21foundation · easy
In the context of AI fundamentals, what is the primary role of a neural network in a machine learning system?
Select an answer first
22expert · hard
A red team is testing an AI-based intrusion detection system (IDS) that uses a neural network to classify network flows. The team wants to evaluate whether an attacker could evade detection without causing noticeable changes to the traffic. Which approach best simulates a realistic evasion attempt while keeping the test within scope?
Select an answer first
23foundation · easy
Which method is commonly used in adversary emulation to test AI-driven security systems?
Select an answer first
24application · medium
A security team deploys an AI-based email gateway that filters phishing messages. The model was trained on a public dataset of emails from 2020–2022. After deployment, the gateway's false-negative rate for phishing increases sharply over two months. The team confirms the model's code and infrastructure are unchanged. Which action best addresses the most likely root cause?
Select an answer first
25expert · medium
A security team is evaluating an AI-based intrusion detection system. They have a test dataset with 10,000 benign samples and 100 malicious samples. The model flags 150 samples as malicious, of which 50 are actually malicious. What is the model's precision for the malicious class?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GASAE

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 GIAC (SANS). “GASAE” is a trademark of its owner, used for identification only.

Artificial Intelligence Fundamentals GASAE Practice Questions (Page 5)

Questions 21–25

In the context of AI fundamentals, what is the primary role of a neural network in a machine learning system?

Which method is commonly used in adversary emulation to test AI-driven security systems?

A security team is evaluating an AI-based intrusion detection system. They have a test dataset with 10,000 benign samples and 100 malicious samples. The model flags 150 samples as malicious, of which 50 are actually malicious. What is the model's precision for the malicious class?

Keep studying — AI and Adversary Emulation