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 54

Question 530

An independent AI auditor reviews this case: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance. Which TWO complementary elements should the auditor seek: an effective corrective control and the most probative independent evidence? Select TWO answers.

  1. Review endpoint inventory, credentials, retained artifacts, disposal approvals, legal holds and restoration paths.
  2. Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.
  5. Separate development, test and production roles with controlled promotion, scoped credentials and audit trails.

Explanation

Correct Answers

B:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. For this situation: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance.

D:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. In this case: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance.

Incorrect Answers

A:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. It does not establish control effectiveness for AI vendor component and supply chain lifecycle.

C:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. For this situation: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance.

E:

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 AI vendor component and supply chain lifecycle.

Question 531

A governance and assurance review uncovers the following: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance. What evidence should the auditor examine FIRST to substantiate or challenge the control owner’s claim? Select ONE best answer.

  1. Assume model weights and datasets require no provenance controls because they are not executable source code.
  2. Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.
  3. Inspect image dependencies, seeds, pipelines, registry signatures, lineage and controlled reruns.
  4. Review endpoint inventory, credentials, retained artifacts, disposal approvals, legal holds and restoration paths.

Explanation

Correct Answer

B:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. This evidence directly tests the asserted control for AI vendor component and supply chain lifecycle: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

Incorrect Answers

A:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. The auditor should instead examine: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

C:

This option concerns Reproducible model training and configuration management, which is a related but different AI governance concern. It does not establish control effectiveness for AI vendor component and supply chain lifecycle.

D:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. For this situation: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance.

Question 532

A risk-based audit engagement examines this operating condition: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance. After management addresses the issue, which result MOST convincingly demonstrates effective remediation? Select ONE best answer.

  1. Production release occurs only after required controls and independent signoffs are demonstrated.
  2. Only approved traceable versions run and vulnerable components receive monitored remediation.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. The retired model cannot serve unauthorized predictions and remaining data is governed by approved retention.

Explanation

Correct Answer

B:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. This is an observable validation result for AI vendor component and supply chain lifecycle, not just management assurance: Only approved traceable versions run and vulnerable components receive monitored remediation.

Incorrect Answers

A:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. For this situation: A company downloads an open-source AI model checkpoint from an unverified mirror without provenance.

C:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. Effective closure requires: Only approved traceable versions run and vulnerable components receive monitored remediation.

D:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. The decisive principle here is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

Question 533

An audit committee asks for a defensible assessment of this situation: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build. 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. Use a decommission plan to disable endpoints, revoke access, retain required records and dispose of data appropriately.
  3. Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.
  4. Assume model weights and datasets require no provenance controls because they are not executable source code.

Explanation

Correct Answer

C:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. This is the proportionate control response to AI vendor component and supply chain lifecycle: Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.

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: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

B:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. The audit must corroborate: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

D:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. In this case: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build.

Question 534

During an AI controls assessment, the team learns that an insurance application imports a model tokenizer dependency that secretly executes unexpected code during build. Which principle BEST explains why the observed practice is a material governance or control concern? Select ONE best answer.

  1. AI lifecycle controls include end-of-life risks; deleting an app icon is not an auditable retirement.
  2. A successful demo establishes feasibility, not sufficient reliability or authorization for real customer impact.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

Explanation

Correct Answer

D:

The acceptable verified outcome is: Only approved traceable versions run and vulnerable components receive monitored remediation. This is the applicable AI audit principle for AI vendor component and supply chain lifecycle: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

Incorrect Answers

A:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. For this situation: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build.

B:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. The decisive principle here is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

C:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. For this situation: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build.

Question 535

An independent AI auditor reviews this case: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build. What evidence should the auditor examine FIRST to substantiate or challenge the control owner’s claim? Select ONE best answer.

  1. Inspect dataset snapshots, feature pipeline commits, record counts, model registry links and training execution manifests.
  2. Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. Inspect image dependencies, seeds, pipelines, registry signatures, lineage and controlled reruns.

Explanation

Correct Answer

B:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. The corrective action is: Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.

Incorrect Answers

A:

This option concerns AI training dataset lineage and reproducibility, which is a related but different AI governance concern. It does not establish control effectiveness for AI vendor component and supply chain lifecycle.

C:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.

D:

This option concerns Reproducible model training and configuration management, which is a related but different AI governance concern. The decisive principle here is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

Question 536

A governance and assurance review uncovers the following: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build. After management addresses the issue, which result MOST convincingly demonstrates effective remediation? Select ONE best answer.

  1. Only approved traceable versions run and vulnerable components receive monitored remediation.
  2. An independent challenger confirms specified safety and performance requirements before go-live.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. Production release occurs only after required controls and independent signoffs are demonstrated.

Explanation

Correct Answer

A:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. Independent audit evidence includes: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

Incorrect Answers

B:

This option concerns Independent validation before model deployment, which is a related but different AI governance concern. The decisive principle here is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

C:

It misses the key control principle: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.

D:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. For this situation: An insurance application imports a model tokenizer dependency that secretly executes unexpected code during build.

Question 537

A risk-based audit engagement examines this operating condition: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration. What is the MOST appropriate risk-based audit recommendation that addresses the root control weakness? Select ONE best answer.

  1. Separate sandbox experimentation from production through constrained access, risk approval and evidence-based readiness criteria.
  2. Create traceable business, human impact, security and quality requirements with measurable release and acceptance gates.
  3. Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.
  4. Assume model weights and datasets require no provenance controls because they are not executable source code.

Explanation

Correct Answer

C:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

Incorrect Answers

A:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

B:

This option concerns AI requirements-to-acceptance traceability, which is a related but different AI governance concern. It does not establish control effectiveness for AI vendor component and supply chain lifecycle.

D:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

Question 538

An audit committee asks for a defensible assessment of this situation: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration. Which principle BEST explains why the observed practice is a material governance or control concern? Select ONE best answer.

  1. AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities.
  2. Assume model weights and datasets require no provenance controls because they are not executable source code.
  3. Developer confidence cannot replace independent assessment for consequential model decisions.
  4. Reproducibility requires more than saving an output weight file or project notebook.

Explanation

Correct Answer

A:

The acceptable verified outcome is: Only approved traceable versions run and vulnerable components receive monitored remediation. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

Incorrect Answers

B:

This attractive shortcut does not resolve the AI vendor component and supply chain lifecycle weakness: Assume model weights and datasets require no provenance controls because they are not executable source code. For this situation: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

C:

This option concerns Independent validation before model deployment, which is a related but different AI governance concern. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

D:

This option concerns Reproducible model training and configuration management, which is a related but different AI governance concern. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

Question 539

During an AI controls assessment, the team learns that a medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration. What evidence should the auditor examine FIRST to substantiate or challenge the control owner’s claim? Select ONE best answer.

  1. Review endpoint inventory, credentials, retained artifacts, disposal approvals, legal holds and restoration paths.
  2. Inspect sandbox isolation, tool permissions, go-live checklist, safety test results and deployment records.
  3. Assume model weights and datasets require no provenance controls because they are not executable source code.
  4. Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

Explanation

Correct Answer

D:

The acceptable verified outcome is: Only approved traceable versions run and vulnerable components receive monitored remediation.

Incorrect Answers

A:

This option concerns Controlled model retirement and data disposition, which is a related but different AI governance concern. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

B:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. For this situation: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

C:

For this situation: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

Question 540

An independent AI auditor reviews this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration. Which TWO complementary elements should the auditor seek: an effective corrective control and the most probative independent evidence? Select TWO answers.

  1. Inspect sandbox isolation, tool permissions, go-live checklist, safety test results and deployment records.
  2. Assume model weights and datasets require no provenance controls because they are not executable source code.
  3. Inventory and authenticate model components, third-party datasets and software dependencies throughout acquisition and updates.
  4. Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.
  5. Separate development, test and production roles with controlled promotion, scoped credentials and audit trails.

Explanation

Correct Answers

C:

The reason it is necessary is: AI supply-chain risk includes pre-trained weights, code, data sources and hosted capabilities. For this situation: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

D:

This evidence directly tests the asserted control for AI vendor component and supply chain lifecycle: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

Incorrect Answers

A:

This option concerns AI prototype experimentation versus production safety gates, which is a related but different AI governance concern. The audit must corroborate: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

B:

The auditor should instead examine: Inspect supplier attestations, component inventory, signed digests, license constraints, SBOM and update notices.

E:

This option concerns AI data and model CI/CD separation, which is a related but different AI governance concern. In this case: A medical assistant depends on external embeddings and cannot identify vulnerable versions or contract expiration.

ISACA Advanced in AI Audit (AAIA) Practice Tests