Why is data quality important in AI systems?

It ensures the model's outputs are accurate and reliable

It reduces the need for data storage

It eliminates all security risks

A company is developing an AI model to predict equipment failures. The data science team wants to use production data that contains sensitive operational details. The security team is concerned about data exposure. What is the most appropriate control to implement during the development phase?

Allow all developers unrestricted access to production data

Delay the project until a secure data environment is built

Prohibit the use of production data in development

Use data masking and access controls on the development environment

A retail company is building a recommendation engine using customer purchase history. The data includes personally identifiable information (PII) such as names and email addresses. The data science team wants to use the data for training. Which practice is most important to implement before training the model?

Remove all purchase history data that is older than one year to reduce the dataset size.

Anonymize or pseudonymize the PII so that the model cannot be used to identify individual customers.

Obtain explicit consent from every customer before using their data for AI training.

Store the data in a separate database with stronger access controls than the production database.

A security analyst is reviewing the risks of an AI system that uses a public cloud API for image recognition. The system processes sensitive images uploaded by users. Which risk is most directly associated with using a third-party AI API?

Data leakage if the API provider stores or uses the uploaded images for model improvement.

The API provider may change the model's behavior without notice, affecting accuracy.

The API may be unavailable during peak usage, causing service disruption.

The AI model may be biased against certain image types.

What is the primary purpose of bias detection in AI systems?

To reduce the size of the training dataset

To ensure the model is fast enough for real-time use

To increase the model's accuracy on all inputs

To identify and correct unfair outcomes that may affect certain groups

GIAC Security Leadership

Domain 2Objective 5

Managing Artificial Intelligence GSLC Practice Questions (Page 7)

Part of the Technical Security Management domain, which makes up ~27% of our current practice bank. GIAC (SANS) does not publish an official question count, but from its 180-minute exam (~70–120 total, ~19–32 in this domain), expect 4–6 from this objective — we provide 61 practice questions to prepare you well beyond it. (estimate)

61questions here

13free pages

12concepts

Questions 31–35

31foundation · easy
Why is data quality important in AI systems?
Select an answer first
32application · medium
A company is developing an AI model to predict equipment failures. The data science team wants to use production data that contains sensitive operational details. The security team is concerned about data exposure. What is the most appropriate control to implement during the development phase?
Select an answer first
33application · medium
A retail company is building a recommendation engine using customer purchase history. The data includes personally identifiable information (PII) such as names and email addresses. The data science team wants to use the data for training. Which practice is most important to implement before training the model?
Select an answer first
34application · medium
A security analyst is reviewing the risks of an AI system that uses a public cloud API for image recognition. The system processes sensitive images uploaded by users. Which risk is most directly associated with using a third-party AI API?
Select an answer first
35foundation · easy
What is the primary purpose of bias detection in AI systems?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to GSLC

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). “GSLC” is a trademark of its owner, used for identification only.