AI Control Library / CF-05: Data Governance & Privacy
Designation: Supplemental | Control types represented: Operational and Technical
Control requirement
Respond to suspected or confirmed AI key compromise through authorized containment, scope assessment, replacement, data protection and recovery. Evaluate exposure already created rather than assuming key replacement reverses it.
Why this matters
A compromised key may affect stored ciphertext, access tokens, wrapped data keys and operational availability. Response needs to identify which protections were lost and which data may remain exposed.
How the control types work together
Operational controls define scope, authority and the actions required when conditions are not met. Technical controls produce evidence or enforce the relevant boundaries. Both must apply to the same deployment and operating conditions.
When it applies
Suspected disclosure, misuse, unauthorized grants or integrity loss involving keys protecting AI information and services.
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
Security incident response coordinates with key custodians, data owners and service operators; designated authorities approve containment and restoration.
Implementation guidance
- Define compromise indicators, reporting channels and emergency authority, preserving key identifiers and relevant audit evidence without spreading secrets.
- Identify affected key versions, grants, assets, backups and exposure periods, distinguishing confirmed facts from uncertainty.
- Contain through revocation, identity restrictions or workload isolation, then replace and re-protect data where appropriate to the threat.
- Verify new trust and recovery conditions, assess required notifications through the established process and document residual exposure.
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 |
|---|---|
| Exercise a compromise scenario | Teams identify affected keys, data and decision authority. |
| Test replacement and containment | Unauthorized capability is addressed without relying only on future-write rotation. |
| Review incident closure | Historical exposure and remaining dependencies are assessed. |
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
- Compromise response procedure
- Scope and audit evidence
- Containment and replacement records
- Data assessment and restoration approval
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
Notify incident response immediately and use approved containment. Coordinate potentially destructive key actions with service owners where feasible, without delaying necessary protection against ongoing compromise.
Example in practice
Hypothetical example: A key-service credential is exposed. Response removes its grants, isolates the affected worker and assesses accessed data; generating a new key alone is not accepted as complete remediation.
Related controls
- CF05-S10: AI Encryption Key Revocation
- CF05-S09: AI Encryption Key Rotation
- CF10-003: AI Incident Response Plans
Framework mappings
- CSA AICM (CEK-19)
- NIST AI RMF (MANAGE 3.1)
- 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 Data Governance & Privacy
Version 1.0