ISACA Advanced in AI Audit AAIA Practice Test 38
Question 370
A public chatbot leaks confidential source snippets, but the response team only resets user passwords. 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?
- Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope
- Revoke agent tools, halt transactions and invoke financial incident coordination
- Link model-provider identifiers to affected services and activate targeted containment
- Track trigger conditions, corrective control, owner and verification after closure
Explanation
Correct Answer
A:Correct. Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope is aimed at the actual failure: ai data-exposure playbook lacks containment of model, retrieval and tool paths. The audit should subsequently obtain containment timeline, query logs and privacy incident assessment to test execution, not merely the existence of a policy.
Incorrect Answers
B:“Revoke agent tools, halt transactions and invoke financial incident coordination” is a control for emergency controls lack a tested stop mechanism for automated action, but the audit finding here is aI data-exposure playbook lacks containment of model, retrieval and tool paths. The relevant measure is “Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope”.
C:“Link model-provider identifiers to affected services and activate targeted containment” is a control for supplier incident process lacks AI service dependency mapping, but the audit finding here is aI data-exposure playbook lacks containment of model, retrieval and tool paths. The relevant measure is “Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope”.
D:“Track trigger conditions, corrective control, owner and verification after closure” is a control for root-cause analysis and recurrence prevention are absent from ai incident management, but the audit finding here is aI data-exposure playbook lacks containment of model, retrieval and tool paths. The relevant measure is “Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope”.
Question 371
In an audit of Incident Response Management Specific to AI, managers say this measure is live: Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope. 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?
- Tool shutdown telemetry, transaction reconciliation and recovery approval
- Dependency registry, scoped incident inventory and vendor notice response
- Recurrent-event analysis and sustained effectiveness checks over later cycles
- Containment timeline, query logs and privacy incident assessment
Explanation
Correct Answer
D:Correct. The proposed records—containment timeline, query logs and privacy incident assessment—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is ai data-exposure playbook lacks containment of model, retrieval and tool paths.
Incorrect Answers
A:“Tool shutdown telemetry, transaction reconciliation and recovery approval” would test emergency controls lack a tested stop mechanism for automated action, not the present safeguard against aI data-exposure playbook lacks containment of model, retrieval and tool paths. The expected audit evidence is “Containment timeline, query logs and privacy incident assessment”.
B:“Dependency registry, scoped incident inventory and vendor notice response” would test supplier incident process lacks AI service dependency mapping, not the present safeguard against aI data-exposure playbook lacks containment of model, retrieval and tool paths. The expected audit evidence is “Containment timeline, query logs and privacy incident assessment”.
C:“Recurrent-event analysis and sustained effectiveness checks over later cycles” would test root-cause analysis and recurrence prevention are absent from ai incident management, not the present safeguard against aI data-exposure playbook lacks containment of model, retrieval and tool paths. The expected audit evidence is “Containment timeline, query logs and privacy incident assessment”.
Question 372
In a review of Incident Response Management Specific to AI, investigators cannot establish the expected control proof: containment timeline, query logs and privacy incident assessment. The underlying event was that a public chatbot leaks confidential source snippets, but the response team only resets user passwords. Which control weakness is the MOST likely cause requiring focused audit work?
- Emergency controls lack a tested stop mechanism for automated action
- Supplier incident process lacks AI service dependency mapping
- AI data-exposure playbook lacks containment of model, retrieval and tool paths
- Root-cause analysis and recurrence prevention are absent from AI incident management
Explanation
Correct Answer
C:Correct. The observed condition specifically indicates ai data-exposure playbook lacks containment of model, retrieval and tool paths. A meaningful audit response is to verify this diagnosis and then recommend disable affected retrieval/tool routes, preserve evidence and assess disclosure scope.
Incorrect Answers
A:“Emergency controls lack a tested stop mechanism for automated action” concerns a separate audit weakness, not the missing assurance described by aI data-exposure playbook lacks containment of model, retrieval and tool paths. The correct cause is “AI data-exposure playbook lacks containment of model, retrieval and tool paths”.
B:“Supplier incident process lacks AI service dependency mapping” concerns a separate audit weakness, not the missing assurance described by aI data-exposure playbook lacks containment of model, retrieval and tool paths. The correct cause is “AI data-exposure playbook lacks containment of model, retrieval and tool paths”.
D:“Root-cause analysis and recurrence prevention are absent from AI incident management” concerns a separate audit weakness, not the missing assurance described by aI data-exposure playbook lacks containment of model, retrieval and tool paths. The correct cause is “AI data-exposure playbook lacks containment of model, retrieval and tool paths”.
Question 373
A poisoned training dataset has already been used to deploy three model versions. 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?
- Route severity-based AI alerts to accountable on-call incident responders
- Identify all impacted builds from lineage and quarantine affected artifacts
- Restore manual review controls or pause consequential decisions during degraded mode
- Trigger circuit breaker, approved fallback and incident classification based on customer harm
Explanation
Correct Answer
B:Correct. Identify all impacted builds from lineage and quarantine affected artifacts is aimed at the actual failure: model incidents do not trace data lineage and affected derivatives. The audit should subsequently obtain dependency graph and verified rollback/retirement of derived models to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Route severity-based AI alerts to accountable on-call incident responders” is a control for aI operational alerts are not integrated into incident severity and on-call ownership, but the audit finding here is model incidents do not trace data lineage and affected derivatives. The relevant measure is “Identify all impacted builds from lineage and quarantine affected artifacts”.
C:“Restore manual review controls or pause consequential decisions during degraded mode” is a control for business-continuity mode disables critical decision safeguards, but the audit finding here is model incidents do not trace data lineage and affected derivatives. The relevant measure is “Identify all impacted builds from lineage and quarantine affected artifacts”.
D:“Trigger circuit breaker, approved fallback and incident classification based on customer harm” is a control for aI incident runbooks do not define fallback and business-continuity response, but the audit finding here is model incidents do not trace data lineage and affected derivatives. The relevant measure is “Identify all impacted builds from lineage and quarantine affected artifacts”.
Question 374
In an audit of Incident Response Management Specific to AI, managers say this measure is live: Identify all impacted builds from lineage and quarantine affected artifacts. 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?
- Dependency graph and verified rollback/retirement of derived models
- Pager traces and tested acknowledgement/resolution targets
- Fail-safe drill results and documented manual approval records
- Time-stamped failover exercise and incident response timeline
Explanation
Correct Answer
A:Correct. The proposed records—dependency graph and verified rollback/retirement of derived models—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is model incidents do not trace data lineage and affected derivatives.
Incorrect Answers
B:“Pager traces and tested acknowledgement/resolution targets” would test aI operational alerts are not integrated into incident severity and on-call ownership, not the present safeguard against model incidents do not trace data lineage and affected derivatives. The expected audit evidence is “Dependency graph and verified rollback/retirement of derived models”.
C:“Fail-safe drill results and documented manual approval records” would test business-continuity mode disables critical decision safeguards, not the present safeguard against model incidents do not trace data lineage and affected derivatives. The expected audit evidence is “Dependency graph and verified rollback/retirement of derived models”.
D:“Time-stamped failover exercise and incident response timeline” would test aI incident runbooks do not define fallback and business-continuity response, not the present safeguard against model incidents do not trace data lineage and affected derivatives. The expected audit evidence is “Dependency graph and verified rollback/retirement of derived models”.
Question 375
In a review of Incident Response Management Specific to AI, investigators cannot establish the expected control proof: dependency graph and verified rollback/retirement of derived models. The underlying event was that a poisoned training dataset has already been used to deploy three model versions. Which control weakness is the MOST likely cause requiring focused audit work?
- AI operational alerts are not integrated into incident severity and on-call ownership
- Business-continuity mode disables critical decision safeguards
- AI incident runbooks do not define fallback and business-continuity response
- Model incidents do not trace data lineage and affected derivatives
Explanation
Correct Answer
D:Correct. The observed condition specifically indicates model incidents do not trace data lineage and affected derivatives. A meaningful audit response is to verify this diagnosis and then recommend identify all impacted builds from lineage and quarantine affected artifacts.
Incorrect Answers
A:“AI operational alerts are not integrated into incident severity and on-call ownership” concerns a separate audit weakness, not the missing assurance described by model incidents do not trace data lineage and affected derivatives. The correct cause is “Model incidents do not trace data lineage and affected derivatives”.
B:“Business-continuity mode disables critical decision safeguards” concerns a separate audit weakness, not the missing assurance described by model incidents do not trace data lineage and affected derivatives. The correct cause is “Model incidents do not trace data lineage and affected derivatives”.
C:“AI incident runbooks do not define fallback and business-continuity response” concerns a separate audit weakness, not the missing assurance described by model incidents do not trace data lineage and affected derivatives. The correct cause is “Model incidents do not trace data lineage and affected derivatives”.
Question 376
An AI agent executes unauthorized financial transactions during an integration fault. 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?
- Isolate generated deliverables, notify counsel and correct affected outputs with audit trail
- Preserve sensor, model, action and operator traces while engaging safety response
- Revoke agent tools, halt transactions and invoke financial incident coordination
- Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope
Explanation
Correct Answer
C:Correct. Revoke agent tools, halt transactions and invoke financial incident coordination is aimed at the actual failure: emergency controls lack a tested stop mechanism for automated action. The audit should subsequently obtain tool shutdown telemetry, transaction reconciliation and recovery approval to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Isolate generated deliverables, notify counsel and correct affected outputs with audit trail” is a control for aI-related inaccurate output is not captured as a reportable quality incident, but the audit finding here is emergency controls lack a tested stop mechanism for automated action. The relevant measure is “Revoke agent tools, halt transactions and invoke financial incident coordination”.
B:“Preserve sensor, model, action and operator traces while engaging safety response” is a control for safety incident evidence and physical consequence monitoring are absent, but the audit finding here is emergency controls lack a tested stop mechanism for automated action. The relevant measure is “Revoke agent tools, halt transactions and invoke financial incident coordination”.
D:“Disable affected retrieval/tool routes, preserve evidence and assess disclosure scope” is a control for aI data-exposure playbook lacks containment of model, retrieval and tool paths, but the audit finding here is emergency controls lack a tested stop mechanism for automated action. The relevant measure is “Revoke agent tools, halt transactions and invoke financial incident coordination”.
Question 377
In an audit of Incident Response Management Specific to AI, managers say this measure is live: Revoke agent tools, halt transactions and invoke financial incident coordination. 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?
- Corrected filings, reviewer approvals and causal analysis
- Tool shutdown telemetry, transaction reconciliation and recovery approval
- Unified event reconstruction and emergency-stop telemetry
- Containment timeline, query logs and privacy incident assessment
Explanation
Correct Answer
B:Correct. The proposed records—tool shutdown telemetry, transaction reconciliation and recovery approval—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is emergency controls lack a tested stop mechanism for automated action.
Incorrect Answers
A:“Corrected filings, reviewer approvals and causal analysis” would test aI-related inaccurate output is not captured as a reportable quality incident, not the present safeguard against emergency controls lack a tested stop mechanism for automated action. The expected audit evidence is “Tool shutdown telemetry, transaction reconciliation and recovery approval”.
C:“Unified event reconstruction and emergency-stop telemetry” would test safety incident evidence and physical consequence monitoring are absent, not the present safeguard against emergency controls lack a tested stop mechanism for automated action. The expected audit evidence is “Tool shutdown telemetry, transaction reconciliation and recovery approval”.
D:“Containment timeline, query logs and privacy incident assessment” would test aI data-exposure playbook lacks containment of model, retrieval and tool paths, not the present safeguard against emergency controls lack a tested stop mechanism for automated action. The expected audit evidence is “Tool shutdown telemetry, transaction reconciliation and recovery approval”.
Question 378
In a review of Incident Response Management Specific to AI, investigators cannot establish the expected control proof: tool shutdown telemetry, transaction reconciliation and recovery approval. The underlying event was that an AI agent executes unauthorized financial transactions during an integration fault. Which control weakness is the MOST likely cause requiring focused audit work?
- Emergency controls lack a tested stop mechanism for automated action
- AI-related inaccurate output is not captured as a reportable quality incident
- Safety incident evidence and physical consequence monitoring are absent
- AI data-exposure playbook lacks containment of model, retrieval and tool paths
Explanation
Correct Answer
A:Correct. The observed condition specifically indicates emergency controls lack a tested stop mechanism for automated action. A meaningful audit response is to verify this diagnosis and then recommend revoke agent tools, halt transactions and invoke financial incident coordination.
Incorrect Answers
B:“AI-related inaccurate output is not captured as a reportable quality incident” concerns a separate audit weakness, not the missing assurance described by emergency controls lack a tested stop mechanism for automated action. The correct cause is “Emergency controls lack a tested stop mechanism for automated action”.
C:“Safety incident evidence and physical consequence monitoring are absent” concerns a separate audit weakness, not the missing assurance described by emergency controls lack a tested stop mechanism for automated action. The correct cause is “Emergency controls lack a tested stop mechanism for automated action”.
D:“AI data-exposure playbook lacks containment of model, retrieval and tool paths” concerns a separate audit weakness, not the missing assurance described by emergency controls lack a tested stop mechanism for automated action. The correct cause is “Emergency controls lack a tested stop mechanism for automated action”.
Question 379
A model degradation alarm fires but operations waits for monthly analytics reporting. 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?
- Link model-provider identifiers to affected services and activate targeted containment
- Track trigger conditions, corrective control, owner and verification after closure
- Identify all impacted builds from lineage and quarantine affected artifacts
- Route severity-based AI alerts to accountable on-call incident responders
Explanation
Correct Answer
D:Correct. Route severity-based AI alerts to accountable on-call incident responders is aimed at the actual failure: ai operational alerts are not integrated into incident severity and on-call ownership. The audit should subsequently obtain pager traces and tested acknowledgement/resolution targets to test execution, not merely the existence of a policy.
Incorrect Answers
A:“Link model-provider identifiers to affected services and activate targeted containment” is a control for supplier incident process lacks AI service dependency mapping, but the audit finding here is aI operational alerts are not integrated into incident severity and on-call ownership. The relevant measure is “Route severity-based AI alerts to accountable on-call incident responders”.
B:“Track trigger conditions, corrective control, owner and verification after closure” is a control for root-cause analysis and recurrence prevention are absent from ai incident management, but the audit finding here is aI operational alerts are not integrated into incident severity and on-call ownership. The relevant measure is “Route severity-based AI alerts to accountable on-call incident responders”.
C:“Identify all impacted builds from lineage and quarantine affected artifacts” is a control for model incidents do not trace data lineage and affected derivatives, but the audit finding here is aI operational alerts are not integrated into incident severity and on-call ownership. The relevant measure is “Route severity-based AI alerts to accountable on-call incident responders”.
Question 380
In an audit of Incident Response Management Specific to AI, managers say this measure is live: Route severity-based AI alerts to accountable on-call incident responders. 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?
- Dependency registry, scoped incident inventory and vendor notice response
- Recurrent-event analysis and sustained effectiveness checks over later cycles
- Pager traces and tested acknowledgement/resolution targets
- Dependency graph and verified rollback/retirement of derived models
Explanation
Correct Answer
C:Correct. The proposed records—pager traces and tested acknowledgement/resolution targets—are directly tied to the asserted safeguard and allow its actual operation to be corroborated. The gap being tested is ai operational alerts are not integrated into incident severity and on-call ownership.
Incorrect Answers
A:“Dependency registry, scoped incident inventory and vendor notice response” would test supplier incident process lacks AI service dependency mapping, not the present safeguard against aI operational alerts are not integrated into incident severity and on-call ownership. The expected audit evidence is “Pager traces and tested acknowledgement/resolution targets”.
B:“Recurrent-event analysis and sustained effectiveness checks over later cycles” would test root-cause analysis and recurrence prevention are absent from ai incident management, not the present safeguard against aI operational alerts are not integrated into incident severity and on-call ownership. The expected audit evidence is “Pager traces and tested acknowledgement/resolution targets”.
D:“Dependency graph and verified rollback/retirement of derived models” would test model incidents do not trace data lineage and affected derivatives, not the present safeguard against aI operational alerts are not integrated into incident severity and on-call ownership. The expected audit evidence is “Pager traces and tested acknowledgement/resolution targets”.