IAPP AIGP Practice Test 88
Question 870
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. The development team must show evidence before promoting the next model version. What is the most directly applicable governance control for this situation?
- Use an enforceable launch gate with documented owners and minimum evidence thresholds.
- Run structured cross-functional causal analysis before changing thresholds or retraining blindly.
- Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path.
- Publish an accurate model card with supported performance, risks and appropriate-use guidance.
Explanation
Correct Answer
C:Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path. Correct. It directly reduces or governs the specific documented risk. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Use an enforceable launch gate with documented owners and minimum evidence thresholds. Not selected. That action treats another control objective, not the specific gap described. It belongs with Release readiness checklist: A release gate is a risk acceptance decision supported by evidence, not a calendar milestone. The present case instead concerns Version deprecation and rollback.
B:Run structured cross-functional causal analysis before changing thresholds or retraining blindly. Not selected. That action treats another control objective, not the specific gap described. It belongs with Root-cause analysis of model failures: AI root-cause analysis must distinguish data drift, code defects, feedback loops and model faults. The present case instead concerns Version deprecation and rollback.
D:Publish an accurate model card with supported performance, risks and appropriate-use guidance. Not selected. That action treats another control objective, not the specific gap described. It belongs with Model cards and limitations: Model cards communicate evidence, limits and proper interpretation to deployment teams. The present case instead concerns Version deprecation and rollback.
Question 871
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. The development team must show evidence before promoting the next model version. What should an independent reviewer test before signing off on this specific risk?
- Generate a controlled shift and confirm alerts lead to triage and documented action.
- Compare candidate and current models on safety and subgroup test suites before promotion.
- Verify remediation closes identified exploits without degrading essential safe capabilities.
- Perform a reversible canary release and confirm the previous approved version can be restored.
Explanation
Correct Answer
D:Perform a reversible canary release and confirm the previous approved version can be restored. Correct. It yields measurable evidence that the case-specific safeguard works. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Generate a controlled shift and confirm alerts lead to triage and documented action. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Monitoring and drift: AI monitoring must measure model behavior and data context, not only infrastructure availability. The present case instead concerns Version deprecation and rollback.
B:Compare candidate and current models on safety and subgroup test suites before promotion. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Scheduled retraining governance: Retraining is a new model change that requires testing, traceability and accountable release. The present case instead concerns Version deprecation and rollback.
C:Verify remediation closes identified exploits without degrading essential safe capabilities. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Red-team exercise design: Adversarial testing probes how a system can fail under purposeful misuse rather than normal usage. The present case instead concerns Version deprecation and rollback.
Question 872
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. The development team must show evidence before promoting the next model version. Select TWO steps that together establish the issue and directly address it.
- Supply accurate instructions, risk information, performance limits and post-market monitoring contacts.
- Map threat actors, exposed interfaces, model assets, data and attack paths.
- Review release artifacts, version pins, rollback dependencies and user communication plans.
- Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path.
Explanation
Correct Answers
C:Review release artifacts, version pins, rollback dependencies and user communication plans. Correct. As the first, evidence-gathering half of the requested pair, this establishes the specific risk boundary. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
D:Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path. Correct. As the second, implementation half of the requested pair, this directly applies the needed safeguard. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Supply accurate instructions, risk information, performance limits and post-market monitoring contacts. Not selected. It addresses a different governance mechanism, not the required assessment-and-control pair. It belongs with Public and downstream disclosure: Developer-to-deployer documentation transfers essential risk information across the AI supply chain. The present case instead concerns Version deprecation and rollback.
B:Map threat actors, exposed interfaces, model assets, data and attack paths. Not selected. It addresses a different governance mechanism, not the required assessment-and-control pair. It belongs with AI security threat modeling: AI security includes attacks against data, models, prompts and inference services. The present case instead concerns Version deprecation and rollback.
Question 873
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. Model assurance reviewers need a documented control before approving continued experimentation. What review step identifies the specific control gap rather than an unrelated AI risk?
- Check audit independence, scope, criteria, sampling and access to underlying evidence.
- Review release artifacts, version pins, rollback dependencies and user communication plans.
- Correlate model, dataset, feature pipeline and upstream service versions with the first failure time.
- Compare live input and outcome distributions with the validated baseline and monitored risk limits.
Explanation
Correct Answer
B:Review release artifacts, version pins, rollback dependencies and user communication plans. Correct. It establishes the relevant factual risk boundary before implementation. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Check audit independence, scope, criteria, sampling and access to underlying evidence. Not selected. That check primarily diagnoses a different governance issue. It belongs with Periodic independent audits: Independent audit contributes objective scrutiny that ordinary self-monitoring cannot replace. The present case instead concerns Version deprecation and rollback.
C:Correlate model, dataset, feature pipeline and upstream service versions with the first failure time. Not selected. That check primarily diagnoses a different governance issue. It belongs with Root-cause analysis of model failures: AI root-cause analysis must distinguish data drift, code defects, feedback loops and model faults. The present case instead concerns Version deprecation and rollback.
D:Compare live input and outcome distributions with the validated baseline and monitored risk limits. Not selected. That check primarily diagnoses a different governance issue. It belongs with Monitoring and drift: AI monitoring must measure model behavior and data context, not only infrastructure availability. The present case instead concerns Version deprecation and rollback.
Question 874
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. Model assurance reviewers need a documented control before approving continued experimentation. Which safeguard most directly treats the described failure mode?
- Create an AI-specific incident category with escalation to business owners and risk management.
- Apply security controls for model APIs, training pipelines, retrieval and privileged operations.
- Review required release artifacts: use-case scope, impact assessment, evaluation, security review and rollback plan.
- Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path.
Explanation
Correct Answer
D:Maintain staged rollout, backward-compatible versioning and an exercised emergency rollback path. Correct. It directly reduces or governs the specific documented risk. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Create an AI-specific incident category with escalation to business owners and risk management. Not selected. That action treats another control objective, not the specific gap described. It belongs with AI incident classification: A system can be operationally available while causing a serious output-quality incident. The present case instead concerns Version deprecation and rollback.
B:Apply security controls for model APIs, training pipelines, retrieval and privileged operations. Not selected. That action treats another control objective, not the specific gap described. It belongs with AI security threat modeling: AI security includes attacks against data, models, prompts and inference services. The present case instead concerns Version deprecation and rollback.
C:Review required release artifacts: use-case scope, impact assessment, evaluation, security review and rollback plan. Not selected. This action relates to Release readiness checklist. A release gate is a risk acceptance decision supported by evidence, not a calendar milestone. For the control decision, it does not resolve the specifically described issue of Version deprecation and rollback in An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. Model assurance reviewers need a documented control before approving continued experimentation.
Question 875
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. Model assurance reviewers need a documented control before approving continued experimentation. Which verification task should support the next approval decision?
- Perform a reversible canary release and confirm the previous approved version can be restored.
- Verify remediation closes identified exploits without degrading essential safe capabilities.
- Verify identified nonconformities have owners, evidence-based closure and deadlines.
- Confirm the integrator can apply the necessary restrictions and escalation procedures.
Explanation
Correct Answer
A:Perform a reversible canary release and confirm the previous approved version can be restored. Correct. It yields measurable evidence that the case-specific safeguard works. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
B:Verify remediation closes identified exploits without degrading essential safe capabilities. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Red-team exercise design: Adversarial testing probes how a system can fail under purposeful misuse rather than normal usage. The present case instead concerns Version deprecation and rollback.
C:Verify identified nonconformities have owners, evidence-based closure and deadlines. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Periodic independent audits: Independent audit contributes objective scrutiny that ordinary self-monitoring cannot replace. The present case instead concerns Version deprecation and rollback.
D:Confirm the integrator can apply the necessary restrictions and escalation procedures. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Public and downstream disclosure: Developer-to-deployer documentation transfers essential risk information across the AI supply chain. The present case instead concerns Version deprecation and rollback.
Question 876
An AI service replaces its model nightly and leaves no supported rollback version when behavior regresses. Model assurance reviewers need a documented control before approving continued experimentation. If the control fails, which immediate response most appropriately limits the documented risk?
- Suspend onboarding where missing disclosures would make safe deployment impossible.
- Escalate unresolved critical findings and restrict risky releases until assurance gaps close.
- Roll back to a known-good model and isolate faulty changes after adverse signals.
- Roll back to the previous validated model if retraining degrades accepted thresholds.
Explanation
Correct Answer
C:Roll back to a known-good model and isolate faulty changes after adverse signals. Correct. It limits further reliance or exposure while the risk remains unresolved. Model lifecycle governance needs controlled retirement and rollback, not irreversible swaps.
Incorrect Answers
A:Suspend onboarding where missing disclosures would make safe deployment impossible. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Public and downstream disclosure: Developer-to-deployer documentation transfers essential risk information across the AI supply chain. The present case instead concerns Version deprecation and rollback.
B:Escalate unresolved critical findings and restrict risky releases until assurance gaps close. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Periodic independent audits: Independent audit contributes objective scrutiny that ordinary self-monitoring cannot replace. The present case instead concerns Version deprecation and rollback.
D:Roll back to the previous validated model if retraining degrades accepted thresholds. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Scheduled retraining governance: Retraining is a new model change that requires testing, traceability and accountable release. The present case instead concerns Version deprecation and rollback.
Question 877
A food retailer pilots an AI freshness predictor, but store connectivity and employee training differ from the laboratory environment. The operational owner must decide whether the current deployment can safely proceed. Which diagnostic task establishes the risk relevant to this described system?
- Compare fine-tuning, prompt design and retrieval on update frequency, provenance, cost and rights.
- Identify knowledge freshness, provenance, source authorization and citation requirements.
- Identify each input modality, potential cross-modal failure and missing evaluation coverage.
- Assess business objectives, data availability, operator skills, network conditions and failure impact at each deployment site.
Explanation
Correct Answer
D:Assess business objectives, data availability, operator skills, network conditions and failure impact at each deployment site. Correct. It establishes the relevant factual risk boundary before implementation. The deployment environment and workforce can create new risk even when the model was sound in development.
Incorrect Answers
A:Compare fine-tuning, prompt design and retrieval on update frequency, provenance, cost and rights. Not selected. That check primarily diagnoses a different governance issue. It belongs with Fine-tuning versus prompt or RAG: Fine-tuning adapts model behavior but is not always suitable for rapidly changing facts. The present case instead concerns Deployment-context feasibility.
B:Identify knowledge freshness, provenance, source authorization and citation requirements. Not selected. That check primarily diagnoses a different governance issue. It belongs with Retrieval augmented generation: RAG can improve grounding but does not eliminate hallucination, leakage or access-control obligations. The present case instead concerns Deployment-context feasibility.
C:Identify each input modality, potential cross-modal failure and missing evaluation coverage. Not selected. That check primarily diagnoses a different governance issue. It belongs with Multimodal model assessment: Multimodal AI adds failure channels that text-only benchmarks cannot substantiate. The present case instead concerns Deployment-context feasibility.
Question 878
A food retailer pilots an AI freshness predictor, but store connectivity and employee training differ from the laboratory environment. The operational owner must decide whether the current deployment can safely proceed. What is the strongest proportionate corrective design decision?
- Pilot a small local model or simple algorithm when it meets safety and accuracy constraints.
- Require multimodal evaluation, input-integrity controls and restrictions on unsupported modality combinations.
- Use controlled retrieval from approved current documents with access filtering and source-aware answers.
- Define a site-readiness gate with infrastructure, workforce and performance acceptance criteria.
Explanation
Correct Answer
D:Define a site-readiness gate with infrastructure, workforce and performance acceptance criteria. Correct. It directly reduces or governs the specific documented risk. The deployment environment and workforce can create new risk even when the model was sound in development.
Incorrect Answers
A:Pilot a small local model or simple algorithm when it meets safety and accuracy constraints. Not selected. That action treats another control objective, not the specific gap described. It belongs with Small versus large model: Model size is one factor in suitability and can increase cost and exposure without adding needed value. The present case instead concerns Deployment-context feasibility.
B:Require multimodal evaluation, input-integrity controls and restrictions on unsupported modality combinations. Not selected. That action treats another control objective, not the specific gap described. It belongs with Multimodal model assessment: Multimodal AI adds failure channels that text-only benchmarks cannot substantiate. The present case instead concerns Deployment-context feasibility.
C:Use controlled retrieval from approved current documents with access filtering and source-aware answers. Not selected. That action treats another control objective, not the specific gap described. It belongs with Retrieval augmented generation: RAG can improve grounding but does not eliminate hallucination, leakage or access-control obligations. The present case instead concerns Deployment-context feasibility.
Question 879
A food retailer pilots an AI freshness predictor, but store connectivity and employee training differ from the laboratory environment. The operational owner must decide whether the current deployment can safely proceed. Which observable evidence best demonstrates the chosen safeguard is working?
- Test corrupted images, accented speech and conflicting text for safe handling and escalation.
- Measure tail latency, operational energy use, robustness and error rates on representative meters.
- Run field pilots at representative stores and review user comprehension and prediction failure rates.
- Test version-changed policy questions and confirm citations, permissions and abstention behavior.
Explanation
Correct Answer
C:Run field pilots at representative stores and review user comprehension and prediction failure rates. Correct. It yields measurable evidence that the case-specific safeguard works. The deployment environment and workforce can create new risk even when the model was sound in development.
Incorrect Answers
A:Test corrupted images, accented speech and conflicting text for safe handling and escalation. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Multimodal model assessment: Multimodal AI adds failure channels that text-only benchmarks cannot substantiate. The present case instead concerns Deployment-context feasibility.
B:Measure tail latency, operational energy use, robustness and error rates on representative meters. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Small versus large model: Model size is one factor in suitability and can increase cost and exposure without adding needed value. The present case instead concerns Deployment-context feasibility.
D:Test version-changed policy questions and confirm citations, permissions and abstention behavior. Not selected. That test would not demonstrate this particular risk has been addressed. It belongs with Retrieval augmented generation: RAG can improve grounding but does not eliminate hallucination, leakage or access-control obligations. The present case instead concerns Deployment-context feasibility.
Question 880
A food retailer pilots an AI freshness predictor, but store connectivity and employee training differ from the laboratory environment. The operational owner must decide whether the current deployment can safely proceed. A material exception remains unresolved. Which containment decision best protects affected parties?
- Disable unvalidated modalities until appropriate testing and oversight are in place.
- Disable unsupported policy advice and provide verified staff assistance during remediation.
- Restore a lightweight validated method when the large model cannot meet edge constraints.
- Limit deployment to sites with validated readiness and manual fallback.
Explanation
Correct Answer
D:Limit deployment to sites with validated readiness and manual fallback. Correct. It limits further reliance or exposure while the risk remains unresolved. The deployment environment and workforce can create new risk even when the model was sound in development.
Incorrect Answers
A:Disable unvalidated modalities until appropriate testing and oversight are in place. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Multimodal model assessment: Multimodal AI adds failure channels that text-only benchmarks cannot substantiate. The present case instead concerns Deployment-context feasibility.
B:Disable unsupported policy advice and provide verified staff assistance during remediation. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Retrieval augmented generation: RAG can improve grounding but does not eliminate hallucination, leakage or access-control obligations. The present case instead concerns Deployment-context feasibility.
C:Restore a lightweight validated method when the large model cannot meet edge constraints. Not selected. That response contains a different failure mode and leaves the stated risk unresolved. It belongs with Small versus large model: Model size is one factor in suitability and can increase cost and exposure without adding needed value. The present case instead concerns Deployment-context feasibility.