IAPP AIGP Sample Questions & Practice Test

The Artificial Intelligence Governance Professional (AIGP) credential is for candidates who need to reason about AI governance, accountability, lifecycle risk, documentation, human oversight, and responsible deployment.

Use these 12 original sample questions for initial self-assessment. They are not official IAPP questions and do not reproduce a live exam.

What this route should test

• AI governance roles, accountability, policy, and oversight
• risk assessment across design, development, deployment, monitoring, and retirement
• transparency, explainability, data stewardship, fairness, and human review
• choosing governance responses without drifting into generic AI hype

Official-source check

Verify current certification names, exam policies, and requirements with the IAPP certification page .

Sample Exam Questions

Question 1

Topic: governance accountability

A company deploys an AI tool that affects customer eligibility decisions. Which governance control is most important before rollout?

• A. Assigning accountable owners, documenting intended use, and approving risk controls
• B. Hiding the model from compliance review
• C. Letting each product team define risk independently
• D. Removing all human review because the model is automated

Best answer: A

Explanation: AI governance starts with accountable ownership, intended-use clarity, risk assessment, and control approval. Automation does not remove organizational responsibility.

Question 2

Topic: model documentation

What is the strongest reason to maintain model cards or similar AI-system documentation?

• A. To prove the model can never fail
• B. To replace all testing
• C. To record purpose, limits, training context, performance, risks, and appropriate use
• D. To avoid stakeholder communication

Best answer: C

Explanation: Documentation helps reviewers understand how the system should be used, where it may fail, and what controls are needed.

Question 3

Topic: human oversight

An AI recommendation is used in a high-impact workflow. What makes human oversight meaningful?

• A. The reviewer sees only a final score with no context
• B. The reviewer can understand, challenge, override, and escalate the recommendation
• C. The reviewer is punished for disagreeing with the system
• D. The reviewer is added after all decisions are final

Best answer: B

Explanation: Oversight is effective only when a human can interpret the recommendation and take action before harm occurs.

Question 4

Topic: data governance

Why is training-data provenance important for AI governance?

• A. It eliminates the need for validation
• B. It makes all personal data anonymous
• C. It is relevant only after an incident
• D. It helps assess rights, representativeness, quality, bias, and permitted use

Best answer: D

Explanation: Provenance supports legal, ethical, and technical review of the data behind an AI system.

Question 5

Topic: fairness risk

A model performs well overall but poorly for one protected or vulnerable group. What is the best governance response?

• A. Ignore the issue because the average metric is good
• B. Publish only the favorable metric
• C. Investigate subgroup performance, root causes, mitigations, and deployment limits
• D. Remove all performance monitoring

Best answer: C

Explanation: Aggregate performance can hide unfair or unsafe outcomes. Governance review should examine subgroup impacts and mitigation options.

Question 6

Topic: lifecycle monitoring

Why does an approved AI system still need post-deployment monitoring?

• A. Data, user behavior, model performance, and risk conditions can change over time
• B. Approval means the system is permanently safe
• C. Monitoring is only needed for failed systems
• D. Monitoring is unrelated to governance

Best answer: A

Explanation: Drift and changing context can affect performance, fairness, and compliance. Governance must continue after launch.

Question 7

Topic: transparency

Which transparency measure is most useful for affected users?

• A. A generic statement that AI is always accurate
• B. A confidential source-code dump
• C. No disclosure unless litigation begins
• D. Clear notice that AI is used, what it supports, and how to seek review or recourse

Best answer: D

Explanation: Transparency should be understandable and actionable. Users need to know when AI affects them and what options exist.

Question 8

Topic: vendor AI risk

An organization buys an AI service from a vendor. Which control is most relevant?

• A. No review is needed because the model is external
• B. Contractual, security, privacy, performance, audit, and change-notice requirements
• C. Letting the vendor define the organization’s risk appetite
• D. Avoiding all documentation

Best answer: B

Explanation: Outsourcing does not outsource accountability. Vendor governance should cover contractual rights, security, data, performance, monitoring, and change management.

Question 9

Topic: risk tiering

Why classify AI systems by risk tier?

• A. To apply governance effort proportionate to potential impact and harm
• B. To avoid reviewing high-impact systems
• C. To make low-risk tools illegal
• D. To remove business ownership

Best answer: A

Explanation: Risk tiering helps focus governance resources where consequences are greater.

Question 10

Topic: incident response

An AI system produces unexpected harmful outputs in production. What should happen first?

• A. Delete logs immediately
• B. Blame the user without review
• C. Preserve evidence, contain impact, notify accountable teams, and follow the incident process
• D. Continue deployment unchanged

Best answer: C

Explanation: AI incidents require evidence preservation, containment, accountability, and structured remediation.

Question 11

Topic: explainability

When is explainability most important?

• A. Only when a model is open source
• B. Never, if the model is complex
• C. Only for marketing copy
• D. When stakeholders need to understand the basis of a material or high-impact decision

Best answer: D

Explanation: Explainability supports oversight, user understanding, auditability, and dispute handling in important decisions.

Question 12

Topic: governance board

What is a useful role for an AI governance committee?

• A. Writing every line of production code
• B. Setting policy, reviewing high-risk use cases, resolving escalations, and monitoring governance outcomes
• C. Eliminating product accountability
• D. Approving all AI use without evidence

Best answer: B

Explanation: Governance committees coordinate risk oversight, policy, escalation, and accountability without replacing operating teams.