Pass Isaca AAIA Exam in First Attempt Easily
Real Isaca AAIA Exam Questions, Accurate & Verified Answers As Experienced in the Actual Test!

Verified by experts
2 products

You save $34.99

AAIA Premium Bundle

  • Premium File 556 Questions & Answers
  • Last Update: Nov 14, 2026
  • Training Course 110 Lectures
$74.99 $109.98

Purchase Individually

  • Premium File

    556 Questions & Answers
    Last Update: Nov 14, 2026

    $76.99
    $69.99
  • Training Course

    110 Lectures

    $43.99
    $39.99

ISACA Advanced in AI Audit AAIA Practice Test 13

Question 120

The AAIA audit team is planning corroboration for this case: A mortgage model passes overall accuracy tests but has materially higher false-rejection rates for one legally protected demographic group. Which records should an auditor examine FIRST to evaluate the relevant control gap?

  1. Inspect stratified metrics, representative labels, sampling uncertainty and fairness review.
  2. Inspect meeting packets, decision register, approvals, exceptions and action follow-ups.
  3. Inspect perturbation test matrix, feature validation, holdout results and operator runbooks.
  4. Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history.

Explanation

Correct Answer

A:

Inspect stratified metrics, representative labels, sampling uncertainty and fairness review. These artifacts are most relevant to evaluating 'Evaluating bias risk across customer groups'. The control under review is: Require justified subgroup performance and fairness analysis with appropriate legal/privacy safeguards and remediation ownership. Expected validated state: Validation reports subgroup outcomes, confidence intervals and any accepted tradeoffs rather than averages alone.

Incorrect Answers

B:

Inspect meeting packets, decision register, approvals, exceptions and action follow-ups. These records are suited to investigating 'Validating governance committee decision evidence', but not the first evidence for 'Evaluating bias risk across customer groups'. Begin with: Inspect stratified metrics, representative labels, sampling uncertainty and fairness review.

C:

Inspect perturbation test matrix, feature validation, holdout results and operator runbooks. These records are suited to investigating 'Testing AI robustness under corrupted but plausible sensor inputs', but not the first evidence for 'Evaluating bias risk across customer groups'. Begin with: Inspect stratified metrics, representative labels, sampling uncertainty and fairness review.

D:

Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history. These records are suited to investigating 'Separating model drift from ordinary process variance', but not the first evidence for 'Evaluating bias risk across customer groups'. Begin with: Inspect stratified metrics, representative labels, sampling uncertainty and fairness review.

Question 121

An ISACA Advanced in AI Audit (AAIA) auditor examines this situation: A procurement risk model deteriorates after vendors shift invoicing formats, yet operations attributes higher false positives to a temporary software outage. Which audit action or control assessment is MOST appropriate?

  1. Assess applicable legal duties by use case, jurisdiction, stakeholder group and deployment role before sign-off.
  2. Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes.
  3. Assess formal retirement gates for shutdown, archival, data retention and access revocation.
  4. Assess prompt-injection threat modeling, trust boundaries and control tests against unauthorized tool and data access.

Explanation

Correct Answer

B:

Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes. Control principle: Changed input distribution and changing relationships to target labels create different model-risk mechanisms. Expected evidence: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality. Real-world boundary: Confirm current applicable law, approved policies, vendor/model version, risk appetite, data permissions, audit standards, business workflow and evidenced operational conditions before adopting an educational example.

Incorrect Answers

A:

Assess applicable legal duties by use case, jurisdiction, stakeholder group and deployment role before sign-off. This addresses the separate audit case 'Evaluating jurisdiction-specific AI legal obligations'. It does not adequately resolve the specific assurance need in 'Separating model drift from ordinary process variance'. The supported focus here is: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes.

C:

Assess formal retirement gates for shutdown, archival, data retention and access revocation. This addresses the separate audit case 'Checking lifecycle coverage for AI model decommissioning'. It does not adequately resolve the specific assurance need in 'Separating model drift from ordinary process variance'. The supported focus here is: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes.

D:

Assess prompt-injection threat modeling, trust boundaries and control tests against unauthorized tool and data access. This addresses the separate audit case 'Evaluating adversarial prompt-injection risk'. It does not adequately resolve the specific assurance need in 'Separating model drift from ordinary process variance'. The supported focus here is: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes.

Question 122

During an independent AAIA assurance review, auditors observe: A procurement risk model deteriorates after vendors shift invoicing formats, yet operations attributes higher false positives to a temporary software outage. Which evidence BEST confirms that the relevant control objective has been met?

  1. Pilot scorecards connect realized benefits, cost and risk to accountable sponsorship and explicit continuation gates.
  2. Cross-tenant negative tests no longer retrieve protected passages and the incident timeline records affected tenants.
  3. Material model failures appear in enterprise risk dashboards with control design, residual risk and accepted treatment.
  4. A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

Explanation

Correct Answer

D:

A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality. This directly substantiates the control objective in 'Separating model drift from ordinary process variance'. Appropriate auditor assessment: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes. Technical and governance rationale: Changed input distribution and changing relationships to target labels create different model-risk mechanisms.

Incorrect Answers

A:

Pilot scorecards connect realized benefits, cost and risk to accountable sponsorship and explicit continuation gates. This would substantiate the different audit case 'Verifying AI portfolio alignment with enterprise strategy', not the key control for 'Separating model drift from ordinary process variance'. The case-specific evidence to obtain instead is: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

B:

Cross-tenant negative tests no longer retrieve protected passages and the incident timeline records affected tenants. This would substantiate the different audit case 'Managing AI prompt-data exposure during a breach', not the key control for 'Separating model drift from ordinary process variance'. The case-specific evidence to obtain instead is: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

C:

Material model failures appear in enterprise risk dashboards with control design, residual risk and accepted treatment. This would substantiate the different audit case 'Integrating AI model risk into enterprise risk registers', not the key control for 'Separating model drift from ordinary process variance'. The case-specific evidence to obtain instead is: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

Question 123

An AI audit team must explain the underlying issue in this case: A procurement risk model deteriorates after vendors shift invoicing formats, yet operations attributes higher false positives to a temporary software outage. Which principle BEST explains why the appropriate audit conclusion follows?

  1. Changed input distribution and changing relationships to target labels create different model-risk mechanisms.
  2. AI program objectives should be testable business outcomes rather than adoption metrics alone.
  3. Audit recommendations should target actual residual risk and avoid blanket measures unconnected to evidence.
  4. Operational risk depends on workflow design, user behavior and downstream response to model output.

Explanation

Correct Answer

A:

Changed input distribution and changing relationships to target labels create different model-risk mechanisms. This is the central audit principle in 'Separating model drift from ordinary process variance'. Suitable review action: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes. Meaningful validation evidence: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

Incorrect Answers

B:

AI program objectives should be testable business outcomes rather than adoption metrics alone. That reasoning concerns 'Verifying AI portfolio alignment with enterprise strategy' and does not explain the material assurance question in 'Separating model drift from ordinary process variance'. The governing mechanism here is: Changed input distribution and changing relationships to target labels create different model-risk mechanisms.

C:

Audit recommendations should target actual residual risk and avoid blanket measures unconnected to evidence. That reasoning concerns 'Setting corrective actions proportional to AI control gaps' and does not explain the material assurance question in 'Separating model drift from ordinary process variance'. The governing mechanism here is: Changed input distribution and changing relationships to target labels create different model-risk mechanisms.

D:

Operational risk depends on workflow design, user behavior and downstream response to model output. That reasoning concerns 'Assessing AI system risk beyond model accuracy' and does not explain the material assurance question in 'Separating model drift from ordinary process variance'. The governing mechanism here is: Changed input distribution and changing relationships to target labels create different model-risk mechanisms.

Question 124

An AI governance or operational control remains inadequate in this scenario: A procurement risk model deteriorates after vendors shift invoicing formats, yet operations attributes higher false positives to a temporary software outage. Which outcome is the MOST direct resulting business or assurance risk?

  1. Inflated testing accuracy can support unsafe deployment of a model that fails on new live claims.
  2. Global deployment may turn one unrecognized model error into a widespread disruption.
  3. Isolated technical logs may conceal cumulative exposure to business decision errors.
  4. Misdiagnosis can drive repeated restarts instead of retraining or policy adjustment.

Explanation

Correct Answer

D:

Misdiagnosis can drive repeated restarts instead of retraining or policy adjustment. This is the most direct risk in 'Separating model drift from ordinary process variance'. Audit response: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes. Independent assurance evidence: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

Incorrect Answers

A:

Inflated testing accuracy can support unsafe deployment of a model that fails on new live claims. This consequence belongs to 'Evaluating supervised fraud labels for leakage' rather than the closest risk in 'Separating model drift from ordinary process variance'. The directly relevant exposure is: Misdiagnosis can drive repeated restarts instead of retraining or policy adjustment.

B:

Global deployment may turn one unrecognized model error into a widespread disruption. This consequence belongs to 'Evaluating canary releases for high-impact AI' rather than the closest risk in 'Separating model drift from ordinary process variance'. The directly relevant exposure is: Misdiagnosis can drive repeated restarts instead of retraining or policy adjustment.

C:

Isolated technical logs may conceal cumulative exposure to business decision errors. This consequence belongs to 'Integrating AI model risk into enterprise risk registers' rather than the closest risk in 'Separating model drift from ordinary process variance'. The directly relevant exposure is: Misdiagnosis can drive repeated restarts instead of retraining or policy adjustment.

Question 125

The AAIA audit team is planning corroboration for this case: A procurement risk model deteriorates after vendors shift invoicing formats, yet operations attributes higher false positives to a temporary software outage. Which records should an auditor examine FIRST to evaluate the relevant control gap?

  1. Review AI roadmap, investment cases, benefits register, risk appetite and stage gates.
  2. Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history.
  3. Inspect model registry, prompt hashes, release audit, traffic routing and rollback test.
  4. Inspect retrieval boundaries, tool scopes, adversarial test results and output logs.

Explanation

Correct Answer

B:

Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history. These artifacts are most relevant to evaluating 'Separating model drift from ordinary process variance'. The control under review is: Evaluate data and concept drift against baselines and investigate corresponding model and business-process changes. Expected validated state: A diagnostic report distinguishes pipeline incidents, input distribution shifts and declining outcome quality.

Incorrect Answers

A:

Review AI roadmap, investment cases, benefits register, risk appetite and stage gates. These records are suited to investigating 'Verifying AI portfolio alignment with enterprise strategy', but not the first evidence for 'Separating model drift from ordinary process variance'. Begin with: Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history.

C:

Inspect model registry, prompt hashes, release audit, traffic routing and rollback test. These records are suited to investigating 'Testing rollback after a harmful model update', but not the first evidence for 'Separating model drift from ordinary process variance'. Begin with: Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history.

D:

Inspect retrieval boundaries, tool scopes, adversarial test results and output logs. These records are suited to investigating 'Evaluating adversarial prompt-injection risk', but not the first evidence for 'Separating model drift from ordinary process variance'. Begin with: Inspect feature distributions, label outcomes, pipeline logs, stability indices and retraining history.

Question 126

An ISACA Advanced in AI Audit (AAIA) auditor examines this situation: Five business applications rely on one external language model and management assesses each app as independent for business continuity. Which audit action or control assessment is MOST appropriate?

  1. Assess content licenses, approved use rights and third-party IP exposure before accepting training data provenance.
  2. Assess impact measures, content safety controls and feedback bias in operational oversight.
  3. Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change.
  4. Map AI-specific threats to shared enterprise controls, document gaps and add focused treatment where existing safeguards fail.

Explanation

Correct Answer

C:

Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change. Control principle: Common upstream services create correlated model risks despite separate downstream application teams. Expected evidence: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives. Real-world boundary: Confirm current applicable law, approved policies, vendor/model version, risk appetite, data permissions, audit standards, business workflow and evidenced operational conditions before adopting an educational example.

Incorrect Answers

A:

Assess content licenses, approved use rights and third-party IP exposure before accepting training data provenance. This addresses the separate audit case 'Checking intellectual-property rights for AI training assets'. It does not adequately resolve the specific assurance need in 'Assessing correlated failure across shared foundation models'. The supported focus here is: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change.

B:

Assess impact measures, content safety controls and feedback bias in operational oversight. This addresses the separate audit case 'Checking oversight of feedback loops in recommender systems'. It does not adequately resolve the specific assurance need in 'Assessing correlated failure across shared foundation models'. The supported focus here is: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change.

D:

Map AI-specific threats to shared enterprise controls, document gaps and add focused treatment where existing safeguards fail. This addresses the separate audit case 'Comparing emerging AI risks to existing control capability'. It does not adequately resolve the specific assurance need in 'Assessing correlated failure across shared foundation models'. The supported focus here is: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change.

Question 127

During an independent AAIA assurance review, auditors observe: Five business applications rely on one external language model and management assesses each app as independent for business continuity. Which evidence BEST confirms that the relevant control objective has been met?

  1. Validation sign-off and audit finding closure are performed by appropriately independent accountable roles.
  2. Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.
  3. A backtest demonstrates justified thresholds for both products with risk-owner approval.
  4. Adversarial ticket tests cannot override approved instructions or exfiltrate protected case files.

Explanation

Correct Answer

B:

Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives. This directly substantiates the control objective in 'Assessing correlated failure across shared foundation models'. Appropriate auditor assessment: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change. Technical and governance rationale: Common upstream services create correlated model risks despite separate downstream application teams.

Incorrect Answers

A:

Validation sign-off and audit finding closure are performed by appropriately independent accountable roles. This would substantiate the different audit case 'Separating independent assurance from model operation', not the key control for 'Assessing correlated failure across shared foundation models'. The case-specific evidence to obtain instead is: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

C:

A backtest demonstrates justified thresholds for both products with risk-owner approval. This would substantiate the different audit case 'Reviewing calibration of fraud model thresholds by product', not the key control for 'Assessing correlated failure across shared foundation models'. The case-specific evidence to obtain instead is: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

D:

Adversarial ticket tests cannot override approved instructions or exfiltrate protected case files. This would substantiate the different audit case 'Evaluating adversarial prompt-injection risk', not the key control for 'Assessing correlated failure across shared foundation models'. The case-specific evidence to obtain instead is: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

Question 128

An AI audit team must explain the underlying issue in this case: Five business applications rely on one external language model and management assesses each app as independent for business continuity. Which principle BEST explains why the appropriate audit conclusion follows?

  1. Model performance metrics are meaningful only against the modality, labels and deployment conditions being audited.
  2. Controls that operated once may cease to operate as model versions and business conditions change.
  3. Common controls remain useful but need testing against novel model-specific failure mechanisms.
  4. Common upstream services create correlated model risks despite separate downstream application teams.

Explanation

Correct Answer

D:

Common upstream services create correlated model risks despite separate downstream application teams. This is the central audit principle in 'Assessing correlated failure across shared foundation models'. Suitable review action: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change. Meaningful validation evidence: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

Incorrect Answers

A:

Model performance metrics are meaningful only against the modality, labels and deployment conditions being audited. That reasoning concerns 'Comparing computer-vision validation with text benchmarks' and does not explain the material assurance question in 'Assessing correlated failure across shared foundation models'. The governing mechanism here is: Common upstream services create correlated model risks despite separate downstream application teams.

B:

Controls that operated once may cease to operate as model versions and business conditions change. That reasoning concerns 'Testing controls at different points in AI model lifecycles' and does not explain the material assurance question in 'Assessing correlated failure across shared foundation models'. The governing mechanism here is: Common upstream services create correlated model risks despite separate downstream application teams.

C:

Common controls remain useful but need testing against novel model-specific failure mechanisms. That reasoning concerns 'Comparing emerging AI risks to existing control capability' and does not explain the material assurance question in 'Assessing correlated failure across shared foundation models'. The governing mechanism here is: Common upstream services create correlated model risks despite separate downstream application teams.

Question 129

An AI governance or operational control remains inadequate in this scenario: Five business applications rely on one external language model and management assesses each app as independent for business continuity. Which outcome is the MOST direct resulting business or assurance risk?

  1. Misleading reasons undermine customer contestability and may create compliance exposure.
  2. A technically effective classifier may automate an inappropriate or legally restricted decision.
  3. Stale recommendations may route emergency vehicles to inaccessible roads after an incident.
  4. One provider incident can simultaneously interrupt customer service, compliance, reporting and operations.

Explanation

Correct Answer

D:

One provider incident can simultaneously interrupt customer service, compliance, reporting and operations. This is the most direct risk in 'Assessing correlated failure across shared foundation models'. Audit response: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change. Independent assurance evidence: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

Incorrect Answers

A:

Misleading reasons undermine customer contestability and may create compliance exposure. This consequence belongs to 'Validating explainability for a credit-scoring model' rather than the closest risk in 'Assessing correlated failure across shared foundation models'. The directly relevant exposure is: One provider incident can simultaneously interrupt customer service, compliance, reporting and operations.

B:

A technically effective classifier may automate an inappropriate or legally restricted decision. This consequence belongs to 'Evaluating AI use-case entry criteria before development' rather than the closest risk in 'Assessing correlated failure across shared foundation models'. The directly relevant exposure is: One provider incident can simultaneously interrupt customer service, compliance, reporting and operations.

C:

Stale recommendations may route emergency vehicles to inaccessible roads after an incident. This consequence belongs to 'Determining whether model fallbacks reduce harm' rather than the closest risk in 'Assessing correlated failure across shared foundation models'. The directly relevant exposure is: One provider incident can simultaneously interrupt customer service, compliance, reporting and operations.

Question 130

The AAIA audit team is planning corroboration for this case: Five business applications rely on one external language model and management assesses each app as independent for business continuity. Which records should an auditor examine FIRST to evaluate the relevant control gap?

  1. Inspect supplier assessment, DPIA, model card, DPA, subcontractor list and signed purchase orders.
  2. Inspect complete event population, declined cases, log gaps and extraction filters.
  3. Inspect dependency graph, SLAs, version controls, fallback testing and concentration limits.
  4. Review vendor assurance, model facts, service agreement, test samples and risk treatment sign-off.

Explanation

Correct Answer

C:

Inspect dependency graph, SLAs, version controls, fallback testing and concentration limits. These artifacts are most relevant to evaluating 'Assessing correlated failure across shared foundation models'. The control under review is: Evaluate dependency concentration and aggregate exposure to a shared model/provider outage, update or safety behavior change. Expected validated state: Enterprise scenarios demonstrate how all five functions fail or degrade together and identify viable alternatives.

Incorrect Answers

A:

Inspect supplier assessment, DPIA, model card, DPA, subcontractor list and signed purchase orders. These records are suited to investigating 'Examining vendor AI procurement entry gates', but not the first evidence for 'Assessing correlated failure across shared foundation models'. Begin with: Inspect dependency graph, SLAs, version controls, fallback testing and concentration limits.

B:

Inspect complete event population, declined cases, log gaps and extraction filters. These records are suited to investigating 'Assessing vendor model output sampling bias', but not the first evidence for 'Assessing correlated failure across shared foundation models'. Begin with: Inspect dependency graph, SLAs, version controls, fallback testing and concentration limits.

D:

Review vendor assurance, model facts, service agreement, test samples and risk treatment sign-off. These records are suited to investigating 'Assessing unquantified residual risk in vendor models', but not the first evidence for 'Assessing correlated failure across shared foundation models'. Begin with: Inspect dependency graph, SLAs, version controls, fallback testing and concentration limits.

ISACA Advanced in AI Audit (AAIA) Practice Tests

Purchase Individually

  • Premium File

    556 Questions & Answers
    Last Update: Nov 14, 2026

    $76.99
    $69.99
  • Training Course

    110 Lectures

    $43.99
    $39.99

SPECIAL OFFER: GET 10% OFF
This is ONE TIME OFFER

You save
10%

Enter Your Email Address to Receive Your 10% Off Discount Code

SPECIAL OFFER: GET 10% OFF

You save
10%

Use Discount Code:

A confirmation link was sent to your e-mail.

Please check your mailbox for a message from support@examlabs.com and follow the directions.

Download Free Demo of VCE Exam Simulator

Experience Avanset VCE Exam Simulator for yourself.

Simply submit your email address below to get started with our interactive software demo of your free trial.

  • Realistic exam simulation and exam editor with preview functions
  • Whole exam in a single file with several different question types
  • Customizable exam-taking mode & detailed score reports