An auditor begins reviewing a new AI use case. Which artifact provides the best testable starting point for the audit?

Documented control objectives linked to the use case risks

A model training dataset has unclear provenance. Which evidence most directly supports an audit conclusion about data governance?

A data lineage record with ownership, quality checks, and approval history

A production model shows performance drift. Which combination best demonstrates that monitoring controls are operating effectively?

Approved thresholds, monitoring logs, and documented escalation decisions

Advanced in AI Audit (AAIA) Practice Test: 3 Free Questions

Exam-specific sample

Questions, answers, and option-by-option explanations

Each explanation identifies why the correct option fits and why the distractors do not.

Easy AI governance and audit planning

Question 1

An auditor begins reviewing a new AI use case. Which artifact provides the best testable starting point for the audit?
1. Documented control objectives linked to the use case risks
2. A list of model features preferred by developers
3. The project launch announcement
4. A forecast of the system cost savings
Correct answer: A. Documented control objectives linked to the use case risks
Review option explanations
- A. Correct. Controls tied to identified risks establish criteria that can be tested and evidenced.
- B. Feature preferences do not establish governance or control requirements.
- C. A launch announcement does not define an auditable control environment.
- D. Cost projections may be useful context but do not establish audit criteria.
Medium AI operations and data management

Question 2

A model training dataset has unclear provenance. Which evidence most directly supports an audit conclusion about data governance?
1. A data lineage record with ownership, quality checks, and approval history
2. A developer statement that the data is widely used
3. A model accuracy score from one test set
4. A list of cloud service invoices
Correct answer: A. A data lineage record with ownership, quality checks, and approval history
Review option explanations
- A. Correct. Lineage, ownership, quality, and approval evidence address whether governed data was used.
- B. An unsupported statement is not sufficient audit evidence.
- C. Accuracy alone does not prove governed sourcing or authorized processing.
- D. Invoices show spending, not data lineage or quality governance.
Hard AI monitoring and audit evidence

Question 3

A production model shows performance drift. Which combination best demonstrates that monitoring controls are operating effectively?
1. Approved thresholds, monitoring logs, and documented escalation decisions
2. A promise to retrain the model next quarter
3. A single dashboard screenshot without a date
4. A comparison with an unrelated model
Correct answer: A. Approved thresholds, monitoring logs, and documented escalation decisions
Review option explanations
- A. Correct. Defined thresholds, recurring evidence, and a response record show a control operating over time.
- B. A future intention does not prove current monitoring.
- C. One undated screenshot does not demonstrate a repeatable control.
- D. An unrelated comparison does not evidence the affected model monitoring process.

More

Advanced in AI Audit (AAIA) Practice Test: 3 Free Questions

Questions, answers, and option-by-option explanations

Question 1

Question 2

Question 3

Continue with the full exam-prep cluster

How can I help?