ISACA Advanced in AI Audit AAIA Practice Test 45
Question 440
In a review of Audit Evidence Collection Techniques, investigators cannot establish the expected control proof: read-only export hashes, log lineage and reconciled report totals. The underlying event was that an auditor receives screenshots of an AI ethics dashboard with no dates or underlying source records. Which control weakness is the MOST likely cause requiring focused audit work?
- Third-party assurance scope does not cover the actual AI service used
- Audit collection breaches confidentiality and least-privilege safeguards
- Evidence provenance and completeness cannot be established
- Source-to-report completeness cannot be substantiated
Explanation
Correct Answer
C:Correct. The observed condition specifically indicates evidence provenance and completeness cannot be established. A meaningful audit response is to verify this diagnosis and then recommend obtain timestamped native exports and reconcile to authoritative system logs.
Incorrect Answers
A:“Third-party assurance scope does not cover the actual AI service used” concerns a separate audit weakness, not the missing assurance described by evidence provenance and completeness cannot be established. The correct cause is “Evidence provenance and completeness cannot be established”.
B:“Audit collection breaches confidentiality and least-privilege safeguards” concerns a separate audit weakness, not the missing assurance described by evidence provenance and completeness cannot be established. The correct cause is “Evidence provenance and completeness cannot be established”.
D:“Source-to-report completeness cannot be substantiated” concerns a separate audit weakness, not the missing assurance described by evidence provenance and completeness cannot be established. The correct cause is “Evidence provenance and completeness cannot be established”.
Question 441
A model team submits a model card but cannot show the exact configuration used in production. From an AI audit perspective, which corrective action is the MOST direct response to the weakness described, rather than a useful control for some different AI risk?
- Use approved restricted evidence environments with minimization and access logs
- Correlate signed model registry versions with runtime configuration and approvals
- Capture privacy-safe decision lineage sufficient for controlled replay
- Collect external gateway logs and independently execute negative control tests
Explanation
Correct Answer
B:Correct. Correlate signed model registry versions with runtime configuration and approvals is aimed at the actual failure: design documentation is not linked to deployed model artifacts. The audit should subsequently obtain runtime digest, registry metadata and deployment attestations to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Use approved restricted evidence environments with minimization and access logs” is a control for audit collection breaches confidentiality and least-privilege safeguards, but the audit finding here is design documentation is not linked to deployed model artifacts. The relevant measure is “Correlate signed model registry versions with runtime configuration and approvals”.
C:“Capture privacy-safe decision lineage sufficient for controlled replay” is a control for evidence capture omits input, feature, model version and inference context, but the audit finding here is design documentation is not linked to deployed model artifacts. The relevant measure is “Correlate signed model registry versions with runtime configuration and approvals”.
D:“Collect external gateway logs and independently execute negative control tests” is a control for self-generated attestation is not independent evidence of control effectiveness, but the audit finding here is design documentation is not linked to deployed model artifacts. The relevant measure is “Correlate signed model registry versions with runtime configuration and approvals”.
Question 442
In an audit of Audit Evidence Collection Techniques, managers say this measure is live: Correlate signed model registry versions with runtime configuration and approvals. They claim it corrected a material audit gap. What direct operating evidence would allow the auditor to substantiate that particular claim across a defensible sample?
- Runtime digest, registry metadata and deployment attestations
- Controlled export records, DLP events and access approvals
- Replay record linking inputs, model digest, decision time and postprocessing
- Gateway block evidence and independent test results mapped to the control
Explanation
Correct Answer
A:Correct. The proposed records—runtime digest, registry metadata and deployment attestations—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is design documentation is not linked to deployed model artifacts.
Incorrect Answers
B:“Controlled export records, DLP events and access approvals” would test audit collection breaches confidentiality and least-privilege safeguards, not the present safeguard against design documentation is not linked to deployed model artifacts. The expected audit evidence is “Runtime digest, registry metadata and deployment attestations”.
C:“Replay record linking inputs, model digest, decision time and postprocessing” would test evidence capture omits input, feature, model version and inference context, not the present safeguard against design documentation is not linked to deployed model artifacts. The expected audit evidence is “Runtime digest, registry metadata and deployment attestations”.
D:“Gateway block evidence and independent test results mapped to the control” would test self-generated attestation is not independent evidence of control effectiveness, not the present safeguard against design documentation is not linked to deployed model artifacts. The expected audit evidence is “Runtime digest, registry metadata and deployment attestations”.
Question 443
In a review of Audit Evidence Collection Techniques, investigators cannot establish the expected control proof: runtime digest, registry metadata and deployment attestations. The underlying event was that a model team submits a model card but cannot show the exact configuration used in production. Which control weakness is the MOST likely cause requiring focused audit work?
- Audit collection breaches confidentiality and least-privilege safeguards
- Evidence capture omits input, feature, model version and inference context
- Self-generated attestation is not independent evidence of control effectiveness
- Design documentation is not linked to deployed model artifacts
Explanation
Correct Answer
D:Correct. The observed condition specifically indicates design documentation is not linked to deployed model artifacts. A meaningful audit response is to verify this diagnosis and then recommend correlate signed model registry versions with runtime configuration and approvals.
Incorrect Answers
A:“Audit collection breaches confidentiality and least-privilege safeguards” concerns a separate audit weakness, not the missing assurance described by design documentation is not linked to deployed model artifacts. The correct cause is “Design documentation is not linked to deployed model artifacts”.
B:“Evidence capture omits input, feature, model version and inference context” concerns a separate audit weakness, not the missing assurance described by design documentation is not linked to deployed model artifacts. The correct cause is “Design documentation is not linked to deployed model artifacts”.
C:“Self-generated attestation is not independent evidence of control effectiveness” concerns a separate audit weakness, not the missing assurance described by design documentation is not linked to deployed model artifacts. The correct cause is “Design documentation is not linked to deployed model artifacts”.
Question 444
A supplier provides a generic assurance report that excludes the hosted inference service. From an AI audit perspective, which corrective action is the MOST direct response to the weakness described, rather than a useful control for some different AI risk?
- Capture privacy-safe decision lineage sufficient for controlled replay
- Obtain raw-to-processed lineage and reconcile dropped, transformed and sampled records
- Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence
- Obtain timestamped native exports and reconcile to authoritative system logs
Explanation
Correct Answer
C:Correct. Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence is aimed at the actual failure: third-party assurance scope does not cover the actual ai service used. The audit should subsequently obtain control mapping, report scope analysis and supplier gap responses to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Capture privacy-safe decision lineage sufficient for controlled replay” is a control for evidence capture omits input, feature, model version and inference context, but the audit finding here is third-party assurance scope does not cover the actual AI service used. The relevant measure is “Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence”.
B:“Obtain raw-to-processed lineage and reconcile dropped, transformed and sampled records” is a control for source-to-report completeness cannot be substantiated, but the audit finding here is third-party assurance scope does not cover the actual AI service used. The relevant measure is “Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence”.
D:“Obtain timestamped native exports and reconcile to authoritative system logs” is a control for evidence provenance and completeness cannot be established, but the audit finding here is third-party assurance scope does not cover the actual AI service used. The relevant measure is “Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence”.
Question 445
In an audit of Audit Evidence Collection Techniques, managers say this measure is live: Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence. They claim it corrected a material audit gap. What direct operating evidence would allow the auditor to substantiate that particular claim across a defensible sample?
- Replay record linking inputs, model digest, decision time and postprocessing
- Control mapping, report scope analysis and supplier gap responses
- ETL execution trace and count reconciliation across pipeline stages
- Read-only export hashes, log lineage and reconciled report totals
Explanation
Correct Answer
B:Correct. The proposed records—control mapping, report scope analysis and supplier gap responses—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is third-party assurance scope does not cover the actual ai service used.
Incorrect Answers
A:“Replay record linking inputs, model digest, decision time and postprocessing” would test evidence capture omits input, feature, model version and inference context, not the present safeguard against third-party assurance scope does not cover the actual AI service used. The expected audit evidence is “Control mapping, report scope analysis and supplier gap responses”.
C:“ETL execution trace and count reconciliation across pipeline stages” would test source-to-report completeness cannot be substantiated, not the present safeguard against third-party assurance scope does not cover the actual AI service used. The expected audit evidence is “Control mapping, report scope analysis and supplier gap responses”.
D:“Read-only export hashes, log lineage and reconciled report totals” would test evidence provenance and completeness cannot be established, not the present safeguard against third-party assurance scope does not cover the actual AI service used. The expected audit evidence is “Control mapping, report scope analysis and supplier gap responses”.
Question 446
In a review of Audit Evidence Collection Techniques, investigators cannot establish the expected control proof: control mapping, report scope analysis and supplier gap responses. The underlying event was that a supplier provides a generic assurance report that excludes the hosted inference service. Which control weakness is the MOST likely cause requiring focused audit work?
- Third-party assurance scope does not cover the actual AI service used
- Evidence capture omits input, feature, model version and inference context
- Source-to-report completeness cannot be substantiated
- Evidence provenance and completeness cannot be established
Explanation
Correct Answer
A:Correct. The observed condition specifically indicates third-party assurance scope does not cover the actual ai service used. A meaningful audit response is to verify this diagnosis and then recommend trace assurance report boundaries to the contracted service and bridge gaps with direct evidence.
Incorrect Answers
B:“Evidence capture omits input, feature, model version and inference context” concerns a separate audit weakness, not the missing assurance described by third-party assurance scope does not cover the actual AI service used. The correct cause is “Third-party assurance scope does not cover the actual AI service used”.
C:“Source-to-report completeness cannot be substantiated” concerns a separate audit weakness, not the missing assurance described by third-party assurance scope does not cover the actual AI service used. The correct cause is “Third-party assurance scope does not cover the actual AI service used”.
D:“Evidence provenance and completeness cannot be established” concerns a separate audit weakness, not the missing assurance described by third-party assurance scope does not cover the actual AI service used. The correct cause is “Third-party assurance scope does not cover the actual AI service used”.
Question 447
An auditor downloads production prompts containing client secrets to a personal workstation. From an AI audit perspective, which corrective action is the MOST direct response to the weakness described, rather than a useful control for some different AI risk?
- Obtain raw-to-processed lineage and reconcile dropped, transformed and sampled records
- Collect external gateway logs and independently execute negative control tests
- Correlate signed model registry versions with runtime configuration and approvals
- Use approved restricted evidence environments with minimization and access logs
Explanation
Correct Answer
D:Correct. Use approved restricted evidence environments with minimization and access logs is aimed at the actual failure: audit collection breaches confidentiality and least-privilege safeguards. The audit should subsequently obtain controlled export records, dlp events and access approvals to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Obtain raw-to-processed lineage and reconcile dropped, transformed and sampled records” is a control for source-to-report completeness cannot be substantiated, but the audit finding here is audit collection breaches confidentiality and least-privilege safeguards. The relevant measure is “Use approved restricted evidence environments with minimization and access logs”.
B:“Collect external gateway logs and independently execute negative control tests” is a control for self-generated attestation is not independent evidence of control effectiveness, but the audit finding here is audit collection breaches confidentiality and least-privilege safeguards. The relevant measure is “Use approved restricted evidence environments with minimization and access logs”.
C:“Correlate signed model registry versions with runtime configuration and approvals” is a control for design documentation is not linked to deployed model artifacts, but the audit finding here is audit collection breaches confidentiality and least-privilege safeguards. The relevant measure is “Use approved restricted evidence environments with minimization and access logs”.
Question 448
In an audit of Audit Evidence Collection Techniques, managers say this measure is live: Use approved restricted evidence environments with minimization and access logs. They claim it corrected a material audit gap. What direct operating evidence would allow the auditor to substantiate that particular claim across a defensible sample?
- ETL execution trace and count reconciliation across pipeline stages
- Gateway block evidence and independent test results mapped to the control
- Controlled export records, DLP events and access approvals
- Runtime digest, registry metadata and deployment attestations
Explanation
Correct Answer
C:Correct. The proposed records—controlled export records, dlp events and access approvals—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is audit collection breaches confidentiality and least-privilege safeguards.
Incorrect Answers
A:“ETL execution trace and count reconciliation across pipeline stages” would test source-to-report completeness cannot be substantiated, not the present safeguard against audit collection breaches confidentiality and least-privilege safeguards. The expected audit evidence is “Controlled export records, DLP events and access approvals”.
B:“Gateway block evidence and independent test results mapped to the control” would test self-generated attestation is not independent evidence of control effectiveness, not the present safeguard against audit collection breaches confidentiality and least-privilege safeguards. The expected audit evidence is “Controlled export records, DLP events and access approvals”.
D:“Runtime digest, registry metadata and deployment attestations” would test design documentation is not linked to deployed model artifacts, not the present safeguard against audit collection breaches confidentiality and least-privilege safeguards. The expected audit evidence is “Controlled export records, DLP events and access approvals”.
Question 449
In a review of Audit Evidence Collection Techniques, investigators cannot establish the expected control proof: controlled export records, dlp events and access approvals. The underlying event was that an auditor downloads production prompts containing client secrets to a personal workstation. Which control weakness is the MOST likely cause requiring focused audit work?
- Source-to-report completeness cannot be substantiated
- Audit collection breaches confidentiality and least-privilege safeguards
- Self-generated attestation is not independent evidence of control effectiveness
- Design documentation is not linked to deployed model artifacts
Explanation
Correct Answer
B:Correct. The observed condition specifically indicates audit collection breaches confidentiality and least-privilege safeguards. A meaningful audit response is to verify this diagnosis and then recommend use approved restricted evidence environments with minimization and access logs.
Incorrect Answers
A:“Source-to-report completeness cannot be substantiated” concerns a separate audit weakness, not the missing assurance described by audit collection breaches confidentiality and least-privilege safeguards. The correct cause is “Audit collection breaches confidentiality and least-privilege safeguards”.
C:“Self-generated attestation is not independent evidence of control effectiveness” concerns a separate audit weakness, not the missing assurance described by audit collection breaches confidentiality and least-privilege safeguards. The correct cause is “Audit collection breaches confidentiality and least-privilege safeguards”.
D:“Design documentation is not linked to deployed model artifacts” concerns a separate audit weakness, not the missing assurance described by audit collection breaches confidentiality and least-privilege safeguards. The correct cause is “Audit collection breaches confidentiality and least-privilege safeguards”.
Question 450
Investigators cannot reproduce an AI decision because only the output score was preserved. From an AI audit perspective, which corrective action is the MOST direct response to the weakness described, rather than a useful control for some different AI risk?
- Capture privacy-safe decision lineage sufficient for controlled replay
- Collect external gateway logs and independently execute negative control tests
- Obtain timestamped native exports and reconcile to authoritative system logs
- Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence
Explanation
Correct Answer
A:Correct. Capture privacy-safe decision lineage sufficient for controlled replay is aimed at the actual failure: evidence capture omits input, feature, model version and inference context. The audit should subsequently obtain replay record linking inputs, model digest, decision time and postprocessing to test execution, not merely the existence of a policy.
Incorrect Answers
B:“Collect external gateway logs and independently execute negative control tests” is a control for self-generated attestation is not independent evidence of control effectiveness, but the audit finding here is evidence capture omits input, feature, model version and inference context. The relevant measure is “Capture privacy-safe decision lineage sufficient for controlled replay”.
C:“Obtain timestamped native exports and reconcile to authoritative system logs” is a control for evidence provenance and completeness cannot be established, but the audit finding here is evidence capture omits input, feature, model version and inference context. The relevant measure is “Capture privacy-safe decision lineage sufficient for controlled replay”.
D:“Trace assurance report boundaries to the contracted service and bridge gaps with direct evidence” is a control for third-party assurance scope does not cover the actual AI service used, but the audit finding here is evidence capture omits input, feature, model version and inference context. The relevant measure is “Capture privacy-safe decision lineage sufficient for controlled replay”.