IAPP AIGP Practice Test 21
Question 200
A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A separate assurance review finds the concern is relevant to an expanded user group. What is the most relevant fallback when the promised safeguard cannot be demonstrated?
- Quarantine undisclosed high-risk workflows until an accountable owner completes intake.
- Suspend vendor onboarding or invoke fallback service if critical evidence is unavailable.
- Roll back unauthorized production use and require completion of the missed approval gate.
- Discover actual AI systems, third-party components, data flows and unapproved shadow uses.
Explanation
Correct Answer
B:Suspend vendor onboarding or invoke fallback service if critical evidence is unavailable. Correct. It limits further reliance or exposure while the risk remains unresolved. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
Incorrect Answers
A:Quarantine undisclosed high-risk workflows until an accountable owner completes intake. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with AI inventory and use case intake: An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. The present case instead concerns Third-party procurement assurance.
C:Roll back unauthorized production use and require completion of the missed approval gate. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns Third-party procurement assurance.
D:Discover actual AI systems, third-party components, data flows and unapproved shadow uses. Not selected. This action relates to AI inventory and use case intake. An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. For the contain decision, it does not resolve the specifically described issue of Third-party procurement assurance in A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A separate assurance review finds the concern is relevant to an expanded user group.
Question 201
A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A material change to the operating model has made the earlier risk acceptance incomplete. What review step identifies the specific control gap rather than an unrelated AI risk?
- Compare employee activities with acceptable-use restrictions on sensitive inputs and consequential outputs.
- Detect changes to vendor terms and assess whether they alter approved purposes, transfers or security controls.
- Review vendor evidence gaps, subcontractors, rights, model change notices and incident terms before contract execution.
- Discover actual AI systems, third-party components, data flows and unapproved shadow uses.
Explanation
Correct Answer
C:Review vendor evidence gaps, subcontractors, rights, model change notices and incident terms before contract execution. Correct. It establishes the relevant factual risk boundary before implementation. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
Incorrect Answers
A:Compare employee activities with acceptable-use restrictions on sensitive inputs and consequential outputs. Not selected. That check primarily diagnoses a different governance issue. It belongs with Acceptable AI use policy: An acceptable-use policy translates ethical commitments into actionable limits and review duties. The present case instead concerns Third-party procurement assurance.
B:Detect changes to vendor terms and assess whether they alter approved purposes, transfers or security controls. Not selected. That check primarily diagnoses a different governance issue. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns Third-party procurement assurance.
D:Discover actual AI systems, third-party components, data flows and unapproved shadow uses. Not selected. This action relates to AI inventory and use case intake. An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. For the diagnose decision, it does not resolve the specifically described issue of Third-party procurement assurance in A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A material change to the operating model has made the earlier risk acceptance incomplete.
Question 202
A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A material change to the operating model has made the earlier risk acceptance incomplete. Which safeguard most directly treats the described failure mode?
- Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions.
- Define evidence-based stage gates with accountable sign-offs and objective exit criteria.
- Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention.
- Implement proportionate immutable decision and version records with retention and access controls.
Explanation
Correct Answer
A:Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions. Correct. It directly reduces or governs the specific documented risk. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
Incorrect Answers
B:Define evidence-based stage gates with accountable sign-offs and objective exit criteria. Not selected. That action treats another control objective, not the specific gap described. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns Third-party procurement assurance.
C:Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention. Not selected. That action treats another control objective, not the specific gap described. It belongs with AI incident policy: AI incidents can be harmful without infrastructure downtime and need explicit response criteria. The present case instead concerns Third-party procurement assurance.
D:Implement proportionate immutable decision and version records with retention and access controls. Not selected. That action treats another control objective, not the specific gap described. It belongs with Record keeping and audit trail: Reliable records enable accountability, legal response, bias investigation and meaningful challenge. The present case instead concerns Third-party procurement assurance.
Question 203
A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A material change to the operating model has made the earlier risk acceptance incomplete. Which verification task should support the next approval decision?
- Compare vendor change logs, approvals and updated policy versions to confirm timely action.
- Run a tabletop exercise that tests technical containment, affected-person analysis and regulatory escalation.
- Check vendor deliverables and test a material change notification against contractual timelines.
- Reconcile purchase records, endpoint logs and business declarations against the inventory.
Explanation
Correct Answer
C:Check vendor deliverables and test a material change notification against contractual timelines. Correct. It yields measurable evidence that the case-specific safeguard works. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
Incorrect Answers
A:Compare vendor change logs, approvals and updated policy versions to confirm timely action. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns Third-party procurement assurance.
B:Run a tabletop exercise that tests technical containment, affected-person analysis and regulatory escalation. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with AI incident policy: AI incidents can be harmful without infrastructure downtime and need explicit response criteria. The present case instead concerns Third-party procurement assurance.
D:Reconcile purchase records, endpoint logs and business declarations against the inventory. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with AI inventory and use case intake: An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. The present case instead concerns Third-party procurement assurance.
Question 204
A vendor sells a powerful customer-scoring API but refuses to document training provenance, subprocessors or security testing. A material change to the operating model has made the earlier risk acceptance incomplete. Select TWO complementary actions: first assess the stated concern and then implement its targeted control.
- Inspect the absence of AI-specific discovery, severity, containment and notification triggers.
- Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions.
- Review vendor evidence gaps, subcontractors, rights, model change notices and incident terms before contract execution.
- Require periodic policy review and event-driven reassessment when material external dependencies change.
Explanation
Correct Answers
B:Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions. Correct. As the second, implementation half of the requested pair, this directly applies the needed safeguard. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
C:Review vendor evidence gaps, subcontractors, rights, model change notices and incident terms before contract execution. Correct. As the first, evidence-gathering half of the requested pair, this establishes the specific risk boundary. Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime.
Incorrect Answers
A:Inspect the absence of AI-specific discovery, severity, containment and notification triggers. Not selected. It addresses a different governance mechanism, not the required assessment-and-control pair. It belongs with AI incident policy: AI incidents can be harmful without infrastructure downtime and need explicit response criteria. The present case instead concerns Third-party procurement assurance.
D:Require periodic policy review and event-driven reassessment when material external dependencies change. Not selected. It addresses a different governance mechanism, not the required assessment-and-control pair. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns Third-party procurement assurance.
Question 205
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. The system owner is requesting a documented approval of this existing governance concern. What review step identifies the specific control gap rather than an unrelated AI risk?
- Identify missing design, validation, release and operations assurance gates.
- Discover actual AI systems, third-party components, data flows and unapproved shadow uses.
- Identify what rights attach to input assets, model terms and generated material in the target market.
- Inspect the absence of AI-specific discovery, severity, containment and notification triggers.
Explanation
Correct Answer
D:Inspect the absence of AI-specific discovery, severity, containment and notification triggers. Correct. It establishes the relevant factual risk boundary before implementation. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
A:Identify missing design, validation, release and operations assurance gates. Not selected. That check primarily diagnoses a different governance issue. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns AI incident policy.
B:Discover actual AI systems, third-party components, data flows and unapproved shadow uses. Not selected. That check primarily diagnoses a different governance issue. It belongs with AI inventory and use case intake: An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. The present case instead concerns AI incident policy.
C:Identify what rights attach to input assets, model terms and generated material in the target market. Not selected. That check primarily diagnoses a different governance issue. It belongs with Intellectual property policy: AI intellectual property risks span inputs, model licensing, generated outputs and downstream use. The present case instead concerns AI incident policy.
Question 206
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. The system owner is requesting a documented approval of this existing governance concern. Which safeguard most directly treats the described failure mode?
- Set review and documentation rules for training material rights, output licenses and human clearance.
- Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention.
- Define evidence-based stage gates with accountable sign-offs and objective exit criteria.
- Discover actual AI systems, third-party components, data flows and unapproved shadow uses.
Explanation
Correct Answer
B:Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention. Correct. It directly reduces or governs the specific documented risk. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
A:Set review and documentation rules for training material rights, output licenses and human clearance. Not selected. That action treats another control objective, not the specific gap described. It belongs with Intellectual property policy: AI intellectual property risks span inputs, model licensing, generated outputs and downstream use. The present case instead concerns AI incident policy.
C:Define evidence-based stage gates with accountable sign-offs and objective exit criteria. Not selected. That action treats another control objective, not the specific gap described. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns AI incident policy.
D:Discover actual AI systems, third-party components, data flows and unapproved shadow uses. Not selected. This action relates to AI inventory and use case intake. An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. For the control decision, it does not resolve the specifically described issue of AI incident policy in An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. The system owner is requesting a documented approval of this existing governance concern.
Question 207
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. The system owner is requesting a documented approval of this existing governance concern. Which verification task should support the next approval decision?
- Run a tabletop exercise that tests technical containment, affected-person analysis and regulatory escalation.
- Compare vendor change logs, approvals and updated policy versions to confirm timely action.
- Sample released assets for documented license checks and provenance records.
- Inspect a release record for assessment, independent review, monitoring plan and rollback instructions.
Explanation
Correct Answer
A:Run a tabletop exercise that tests technical containment, affected-person analysis and regulatory escalation. Correct. It yields measurable evidence that the case-specific safeguard works. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
B:Compare vendor change logs, approvals and updated policy versions to confirm timely action. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns AI incident policy.
C:Sample released assets for documented license checks and provenance records. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Intellectual property policy: AI intellectual property risks span inputs, model licensing, generated outputs and downstream use. The present case instead concerns AI incident policy.
D:Inspect a release record for assessment, independent review, monitoring plan and rollback instructions. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns AI incident policy.
Question 208
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. The system owner is requesting a documented approval of this existing governance concern. If the control fails, which immediate response most appropriately limits the documented risk?
- Suspend vendor onboarding or invoke fallback service if critical evidence is unavailable.
- Disable the unsafe output path, preserve logs and initiate incident management immediately.
- Freeze affected processing if a changed term would undermine an essential control.
- Withdraw uncleared materials and seek legal assessment before wider distribution.
Explanation
Correct Answer
B:Disable the unsafe output path, preserve logs and initiate incident management immediately. Correct. It limits further reliance or exposure while the risk remains unresolved. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
A:Suspend vendor onboarding or invoke fallback service if critical evidence is unavailable. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Third-party procurement assurance: Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime. The present case instead concerns AI incident policy.
C:Freeze affected processing if a changed term would undermine an essential control. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns AI incident policy.
D:Withdraw uncleared materials and seek legal assessment before wider distribution. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Intellectual property policy: AI intellectual property risks span inputs, model licensing, generated outputs and downstream use. The present case instead concerns AI incident policy.
Question 209
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. A separate assurance review finds the concern is relevant to an expanded user group. Which initial evidence check most directly addresses the governance concern?
- Detect changes to vendor terms and assess whether they alter approved purposes, transfers or security controls.
- Discover actual AI systems, third-party components, data flows and unapproved shadow uses.
- Identify missing design, validation, release and operations assurance gates.
- Inspect the absence of AI-specific discovery, severity, containment and notification triggers.
Explanation
Correct Answer
D:Inspect the absence of AI-specific discovery, severity, containment and notification triggers. Correct. It establishes the relevant factual risk boundary before implementation. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
A:Detect changes to vendor terms and assess whether they alter approved purposes, transfers or security controls. Not selected. That check primarily diagnoses a different governance issue. It belongs with Policy change control: Governance documentation must be living evidence that follows real systems and contracts. The present case instead concerns AI incident policy.
B:Discover actual AI systems, third-party components, data flows and unapproved shadow uses. Not selected. That check primarily diagnoses a different governance issue. It belongs with AI inventory and use case intake: An accurate AI inventory is the foundation for lifecycle controls, monitoring and incident response. The present case instead concerns AI incident policy.
C:Identify missing design, validation, release and operations assurance gates. Not selected. That check primarily diagnoses a different governance issue. It belongs with Lifecycle gates: Governance policies must attach to concrete lifecycle transitions rather than remain abstract aspirations. The present case instead concerns AI incident policy.
Question 210
An image model unexpectedly discloses internal customer identifiers, but the incident playbook only addresses server outages. A separate assurance review finds the concern is relevant to an expanded user group. Which specific implementation best addresses the identified gap?
- Implement proportionate immutable decision and version records with retention and access controls.
- Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention.
- Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions.
- Publish an approved-use policy with prohibited cases, verification rules, reporting obligations and tool boundaries.
Explanation
Correct Answer
B:Extend incident response procedures to model outputs, data leakage, misuse, reporting and evidence retention. Correct. It directly reduces or governs the specific documented risk. AI incidents can be harmful without infrastructure downtime and need explicit response criteria.
Incorrect Answers
A:Implement proportionate immutable decision and version records with retention and access controls. Not selected. That action treats another control objective, not the specific gap described. It belongs with Record keeping and audit trail: Reliable records enable accountability, legal response, bias investigation and meaningful challenge. The present case instead concerns AI incident policy.
C:Require contractually enforceable assurances, audit rights, change notices, incident cooperation and data-use restrictions. Not selected. That action treats another control objective, not the specific gap described. It belongs with Third-party procurement assurance: Procurement governance must cover the whole AI supply chain, not only the purchase price and uptime. The present case instead concerns AI incident policy.
D:Publish an approved-use policy with prohibited cases, verification rules, reporting obligations and tool boundaries. Not selected. That action treats another control objective, not the specific gap described. It belongs with Acceptable AI use policy: An acceptable-use policy translates ethical commitments into actionable limits and review duties. The present case instead concerns AI incident policy.