To confirm the fairness of AI model decisions, the BEST way to collect reliable evidence during an AI audit is by:

• A.Analyzing system metadata.
• B.Testing the model with a curated sample data set.
• C.Interviewing developers.
• D.Observing the system's interactions with end users.

Comment: *

Name: *

Email: *

Verification: *

An IS auditor for a veterinary clinic was informed that the dog breed categorical variable is necessary for the predictive model. Which of the following introduces the MOST risk?

• A.Data scaling was not utilized.
• B.Clustering was not utilized.
• C.Ordinal label encoding was utilized.
• D.One-hot encoding was utilized.

Comment: *

Name: *

Email: *

Verification: *

An IS auditor finds that an AI model's outputs are not being reviewed. Which of the following would BEST address this risk?

• A.A larger training dataset
• B.A validation process for AI decisions
• C.Regular AI model retraining
• D.Prompt templates

Comment: *

Name: *

Email: *

Verification: *

Which of the following is the MOST important course of action for an organization prior to allowing end users to utilize an AI tool?

• A.Develop an AI policy with guidelines on appropriate use.
• B.Ensure a cybersecurity insurance clause is in place to include the use of AI.
• C.Implement baseline performance metrics.
• D.Determine the impact to the disaster recovery plan (DRP).

Comment: *

Name: *

Email: *

Verification: *

An IS auditor is considering the integration of AI techniques into the audit sampling process. Which of the following BEST enables the auditor to identify high-risk transactions within large data sets for targeted sampling?

• A.Natural language processing (NLP)
• B.Optical character recognition (OCR)
• C.Rule-based analytics
• D.Predictive analytics

Comment: *

Name: *

Email: *

Verification: *