AI Control Library / CF-10: Recourse & Incident Response
Designation: Core | Control types represented: Operational
Control requirement
Triage AI incident reports through prompt intake, protected evidence collection, impact assessment, ownership and initial containment decisions. Preserve uncertainty and connect related cases without requiring confirmed root cause before acting.
Why this matters
An unusual output may be a harmless error, a sign of broader compromise or evidence of actual harm. Triage must establish relevant facts quickly while avoiding both dismissal and unsupported claims.
How the control types work together
Operational controls establish the assigned responsibilities, review processes and response decisions described here. Their effectiveness depends on performing the work and using evidence to guide action, rather than approving a document alone.
When it applies
AI alerts, user complaints, supplier notifications, audit findings and operational reports that may require incident response.
Define the assessed boundary, dependencies and evidence limitations. Tailor depth and review frequency to potential impact, and document why a particular safeguard or exclusion is appropriate to the actual deployment.
Ownership
The triage owner coordinates qualified system, domain, security and privacy responders.
Implementation guidance
- Provide intake and acknowledgement routes, collecting the minimum useful system, version, time, action and impact context.
- Distinguish observed effects from interpretation, preserve relevant records and identify continuing exposure.
- Classify and assign an incident owner, initiating authorized containment and specialist review as appropriate.
- Reassess after new evidence, linking duplicate reports and recording decisions to close, monitor or escalate.
How to verify it
Use representative records and authorized controlled tests. Identify the relevant version, operating conditions and expected result before testing, and retain enough evidence to support the conclusion.
| Test | Expected result |
|---|---|
| Submit a consequential report | It reaches an accountable triage role with sufficient context. |
| Inspect early incident notes | Facts, uncertainty and hypotheses remain distinguishable. |
| Identify ongoing harm | Initial containment is considered before root cause is fully established. |
Record a finding when required results cannot be demonstrated or practice differs materially from the approved conditions. Missing evidence limits the conclusion; it should not be treated as a successful test.
Evidence to retain
- Intake and triage procedure
- Protected report and evidence
- Classification and assignment
- Containment and reassessment decisions
Protect evidence according to its sensitivity, access needs and retention rules. Avoid unnecessary copies of personal information, secrets or confidential material, while preserving the records needed for authorized review.
If the control fails
Escalate credible material harm or uncontrolled activity immediately through the defined route. Preserve evidence without allowing collection to delay necessary protection.
Example in practice
Hypothetical example: A user reports an incorrect account action. Triage checks the tool result and affected accounts rather than inferring execution from the agent’s statement that it completed the task.
Related controls
- CF10-002: AI Incident Category & Severity Classification
- CF04-014: AI Decision Record & Traceability
- CF10-003: AI Incident Response Plans
Framework mappings
- CSA AICM (SEF-06)
- NIST AI RMF (MANAGE 3.1)
- EU AI Act
- ISO 42001 (Cl. 10.2)
Review and maintenance
Review at the interval approved by the control owner and after relevant changes, incidents or failed assessments. Update affected procedures and configuration, address evidence gaps and verify corrective action before closing findings.
Back to Recourse & Incident Response
Version 1.0