AI Control Library / CF-04: Auditability & Evidence
Designation: Supplemental | Control types represented: Operational and Technical
Control requirement
Monitor encryption operations supporting AI services for failures, unexpected bypass, insecure configuration and material gaps in protected data paths. Report findings to accountable operators and verify corrective action.
Why this matters
A policy requiring encryption does not show whether a live connection, storage path or export actually uses the approved protection. Failures can occur at integrations and recovery paths outside the primary model service.
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
Encryption in transit and at rest for AI data, artifacts, logs and supporting services; assess cloud-provider evidence where direct observation is unavailable.
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 encryption service owner maintains monitoring; platform, network and data owners respond to affected-path findings.
Implementation guidance
- Identify encryption boundaries, relevant storage and connections, approved settings and telemetry available for verification.
- Monitor failures, insecure fallback, configuration changes and relevant service health without exposing keys or sensitive payloads.
- Correlate findings with data sensitivity and service impact, routing actionable reports to responsible operators.
- Test representative encryption failures and configuration deviations, then verify restoration of protection on the actual affected path.
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 failed protected connection | The application and monitoring respond according to the approved failure policy. |
| Introduce an unapproved setting in a controlled environment | Detection identifies the affected boundary and owner. |
| Inspect a corrected encryption finding | Evidence confirms protection is restored rather than only the alert being silenced. |
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
- Encryption boundary and telemetry map
- Approved configuration references
- Alerts and response records
- Failure and restoration tests
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 security and affected service owners of material encryption loss or bypass. Block or isolate sensitive transfers where protection is required and investigate exposed information under the incident process.
Example in practice
Hypothetical example: A model export job writes to a new storage location with different encryption settings. Configuration monitoring detects the uncovered path before the team treats the primary database’s encryption as protection for the export.
Related controls
- CF05-017: AI Data Encryption at Rest & in Transit
- CF04-S02: AI Key Lifecycle Audit Logging
- CF09-005: AI Security Monitoring & Alerting
Framework mappings
- CSA AICM (LOG-10)
- NIST AI RMF (MANAGE 4.1)
- ISO 42001 (Cl. 9.1)
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 Auditability & Evidence
Version 1.0