cert
cert-1
cert-2

Pass Isaca AAIA Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
AAIA Exam - Verified By Experts
AAIA Premium Bundle
$14.99

AAIA Premium Bundle

$74.99
$89.98
  • Premium File 556 Questions & Answers. Last update: Nov 08, 2026
  • Training Course 110 Video Lectures
 
$89.98
$74.99
accept 22 downloads in last 7 days
block-screenshots
Purchase Isaca AAIA Exam Training Products Individually
Isaca AAIA Premium File
Premium File 556 Q&A
$76.99$69.99
Isaca AAIA Video Training Course
Training Course 110 Lectures
$21.99 $19.99
Isaca AAIA Practice Test 87

Question 860

An independent AI auditor reviews this case: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed. Which TWO complementary elements should the auditor seek: an effective corrective control and the most probative independent evidence? Select TWO answers.

  1. Treat retrieved text as untrusted data, separate instruction authority and constrain downstream tool access.
  2. Inspect tool manifests, service accounts, approval gates, injection tests and action receipts.
  3. Verify model provenance, storage integrity, executable dependencies and approved component changes.
  4. Assume every pretrained model download from a popular forum is trustworthy without signature or provenance.
  5. Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records.

Explanation

Correct Answers

C:

Independent audit evidence includes: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records. For this situation: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed.

E:

The corrective action is: Verify model provenance, storage integrity, executable dependencies and approved component changes. For this situation: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed.

Incorrect Answers

A:

This option concerns Indirect prompt injection through retrieved content, which is a related but different AI governance concern. In this case: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed.

B:

This option concerns AI tool misuse and excessive agency, which is a related but different AI governance concern. For this situation: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed.

D:

For this situation: A third-party model package includes an unexpected runtime hook but its license and dependencies were never reviewed.

Question 861

A governance and assurance review uncovers the following: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them. What is the MOST appropriate risk-based audit recommendation that addresses the root control weakness? Select ONE best answer.

  1. Treat retrieved text as untrusted data, separate instruction authority and constrain downstream tool access.
  2. Verify model provenance, storage integrity, executable dependencies and approved component changes.
  3. Assume every pretrained model download from a popular forum is trustworthy without signature or provenance.
  4. Scope tools to specific identities and approved operations with pre-execution authorization and transaction caps.

Explanation

Correct Answer

B:

Independent audit evidence includes: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records. The reason it is necessary is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Incorrect Answers

A:

This option concerns Indirect prompt injection through retrieved content, which is a related but different AI governance concern. The decisive principle here is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

C:

This attractive shortcut does not resolve the Foundation model dependency and artifact tampering weakness: Assume every pretrained model download from a popular forum is trustworthy without signature or provenance. In this case: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

D:

This option concerns AI tool misuse and excessive agency, which is a related but different AI governance concern. The decisive principle here is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Question 862

A risk-based audit engagement examines this operating condition: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them. Which principle BEST explains why the observed practice is a material governance or control concern? Select ONE best answer.

  1. Untrusted retrieved documents can manipulate model instructions but must not gain system-level privileges.
  2. AI delivery involves models, data and code, so conventional app deployment approval alone may not cover behavior changes.
  3. Assume every pretrained model download from a popular forum is trustworthy without signature or provenance.
  4. Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Explanation

Correct Answer

D:

Independent audit evidence includes: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records. This is the applicable AI audit principle for Foundation model dependency and artifact tampering: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Incorrect Answers

A:

This option concerns Indirect prompt injection through retrieved content, which is a related but different AI governance concern. For this situation: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

B:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. The decisive principle here is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

C:

This attractive shortcut does not resolve the Foundation model dependency and artifact tampering weakness: Assume every pretrained model download from a popular forum is trustworthy without signature or provenance. For this situation: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

Question 863

An audit committee asks for a defensible assessment of this situation: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them. What evidence should the auditor examine FIRST to substantiate or challenge the control owner’s claim? Select ONE best answer.

  1. Assume every pretrained model download from a popular forum is trustworthy without signature or provenance.
  2. Inspect pipeline roles, artifact signatures, dataset permissions, approval gates and production change logs.
  3. Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records.
  4. Inspect authentication, rate limits, output granularity, query anomalies and abuse response.

Explanation

Correct Answer

C:

The corrective action is: Verify model provenance, storage integrity, executable dependencies and approved component changes. The reason it is necessary is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Incorrect Answers

A:

The auditor should instead examine: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records.

B:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. For this situation: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

D:

This option concerns AI model extraction and excessive query access, which is a related but different AI governance concern. The decisive principle here is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

Question 864

During an AI controls assessment, the team learns that an organization stores production model artifacts in a writable shared bucket and loses track of who replaced them. After management addresses the issue, which result MOST convincingly demonstrates effective remediation? Select ONE best answer.

  1. Only approved corpus changes influence outputs, and unexpected ranking changes trigger investigation.
  2. Malicious document instructions cannot cause prohibited tool actions or confidential-data release.
  3. Assume every pretrained model download from a popular forum is trustworthy without signature or provenance.
  4. Production loads only authenticated approved artifacts and tamper attempts are denied or alerted.

Explanation

Correct Answer

D:

Independent audit evidence includes: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records. In this case: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

Incorrect Answers

A:

This option concerns Poisoned embedding and vector retrieval index, which is a related but different AI governance concern. The decisive principle here is: Model weights, parsers and surrounding runtime code belong in the software and data supply-chain threat model.

B:

This option concerns Indirect prompt injection through retrieved content, which is a related but different AI governance concern. The audit must corroborate: Inspect signatures, hash attestations, artifact ACLs, dependency inventory and rebuild records.

C:

For this situation: An organization stores production model artifacts in a writable shared bucket and loses track of who replaced them.

Question 865

An independent AI auditor reviews this case: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results. What is the MOST appropriate risk-based audit recommendation that addresses the root control weakness? Select ONE best answer.

  1. Enforce point-in-time joins and temporal cutoffs for training and evaluation features.
  2. Scope tools to specific identities and approved operations with pre-execution authorization and transaction caps.
  3. Evaluate unintended inference from model outputs and minimize observable detail and query access.
  4. Assume suppressing direct database access prevents all leakage of private attributes through an API.

Explanation

Correct Answer

C:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. In this case: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results.

Incorrect Answers

A:

This option concerns Feature-store time leakage in model evaluation, which is a related but different AI governance concern. It does not establish control effectiveness for Model inversion and sensitive attribute leakage.

B:

This option concerns AI tool misuse and excessive agency, which is a related but different AI governance concern. It does not establish control effectiveness for Model inversion and sensitive attribute leakage.

D:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API.

Question 866

A governance and assurance review uncovers the following: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results. Which principle BEST explains why the observed practice is a material governance or control concern? Select ONE best answer.

  1. Assume suppressing direct database access prevents all leakage of private attributes through an API.
  2. A model can disclose information statistically through outputs even without returning underlying records verbatim.
  3. AI delivery involves models, data and code, so conventional app deployment approval alone may not cover behavior changes.
  4. Repeated API observations can reveal model behavior or sensitive correlations even without access to the weight files.

Explanation

Correct Answer

B:

Independent audit evidence includes: Inspect output schemas, reconstruction probes, privacy threat model, aggregation and access restrictions. This is the applicable AI audit principle for Model inversion and sensitive attribute leakage: A model can disclose information statistically through outputs even without returning underlying records verbatim.

Incorrect Answers

A:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API. For this situation: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results.

C:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. It does not establish control effectiveness for Model inversion and sensitive attribute leakage.

D:

This option concerns AI model extraction and excessive query access, which is a related but different AI governance concern. It does not establish control effectiveness for Model inversion and sensitive attribute leakage.

Question 867

A risk-based audit engagement examines this operating condition: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results. What evidence should the auditor examine FIRST to substantiate or challenge the control owner’s claim? Select ONE best answer.

  1. Inspect tool manifests, service accounts, approval gates, injection tests and action receipts.
  2. Inspect authentication, rate limits, output granularity, query anomalies and abuse response.
  3. Inspect output schemas, reconstruction probes, privacy threat model, aggregation and access restrictions.
  4. Assume suppressing direct database access prevents all leakage of private attributes through an API.

Explanation

Correct Answer

C:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. This evidence directly tests the asserted control for Model inversion and sensitive attribute leakage: Inspect output schemas, reconstruction probes, privacy threat model, aggregation and access restrictions.

Incorrect Answers

A:

This option concerns AI tool misuse and excessive agency, which is a related but different AI governance concern. For this situation: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results.

B:

This option concerns AI model extraction and excessive query access, which is a related but different AI governance concern. For this situation: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results.

D:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API. The auditor should instead examine: Inspect output schemas, reconstruction probes, privacy threat model, aggregation and access restrictions.

Question 868

An audit committee asks for a defensible assessment of this situation: A face-recognition embedding API allows repeated probes that reconstruct sensitive characteristics from similarity results. After management addresses the issue, which result MOST convincingly demonstrates effective remediation? Select ONE best answer.

  1. Only approved corpus changes influence outputs, and unexpected ranking changes trigger investigation.
  2. Unintended sensitive inference is bounded and logged access follows approved purpose.
  3. Assume suppressing direct database access prevents all leakage of private attributes through an API.
  4. Suspicious extraction patterns are detected and unauthorized systematic probing is constrained.

Explanation

Correct Answer

B:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. This is an observable validation result for Model inversion and sensitive attribute leakage, not just management assurance: Unintended sensitive inference is bounded and logged access follows approved purpose.

Incorrect Answers

A:

This option concerns Poisoned embedding and vector retrieval index, which is a related but different AI governance concern. It does not establish control effectiveness for Model inversion and sensitive attribute leakage.

C:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API. Effective closure requires: Unintended sensitive inference is bounded and logged access follows approved purpose.

D:

This option concerns AI model extraction and excessive query access, which is a related but different AI governance concern. The decisive principle here is: A model can disclose information statistically through outputs even without returning underlying records verbatim.

Question 869

During an AI controls assessment, the team learns that a health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task. What is the MOST appropriate risk-based audit recommendation that addresses the root control weakness? Select ONE best answer.

  1. Separate development, test and production roles with controlled promotion, scoped credentials and audit trails.
  2. Enforce point-in-time joins and temporal cutoffs for training and evaluation features.
  3. Assume suppressing direct database access prevents all leakage of private attributes through an API.
  4. Evaluate unintended inference from model outputs and minimize observable detail and query access.

Explanation

Correct Answer

D:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. In this case: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

Incorrect Answers

A:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. The decisive principle here is: A model can disclose information statistically through outputs even without returning underlying records verbatim.

B:

This option concerns Feature-store time leakage in model evaluation, which is a related but different AI governance concern. The decisive principle here is: A model can disclose information statistically through outputs even without returning underlying records verbatim.

C:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API. In this case: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

Question 870

An independent AI auditor reviews this case: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task. Which TWO complementary elements should the auditor seek: an effective corrective control and the most probative independent evidence? Select TWO answers.

  1. Evaluate unintended inference from model outputs and minimize observable detail and query access.
  2. Rate-limit and monitor queries, restrict unnecessary output detail and evaluate extraction risks proportionate to value.
  3. Inspect pipeline roles, artifact signatures, dataset permissions, approval gates and production change logs.
  4. Assume suppressing direct database access prevents all leakage of private attributes through an API.
  5. Inspect output schemas, reconstruction probes, privacy threat model, aggregation and access restrictions.

Explanation

Correct Answers

A:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. For this situation: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

E:

The reason it is necessary is: A model can disclose information statistically through outputs even without returning underlying records verbatim. The corrective action is: Evaluate unintended inference from model outputs and minimize observable detail and query access.

Incorrect Answers

B:

This option concerns AI model extraction and excessive query access, which is a related but different AI governance concern. In this case: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

C:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. For this situation: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

D:

This attractive shortcut does not resolve the Model inversion and sensitive attribute leakage weakness: Assume suppressing direct database access prevents all leakage of private attributes through an API. For this situation: A health risk model's detailed prediction responses permit inference of protected attributes beyond the intended task.

ISACA Advanced in AI Audit (AAIA) Practice Tests