AI Control Library / CF-12: Security & Abuse Resistance
Designation: Supplemental | Control types represented: Operational and Technical
Control requirement
Protect AI audit logging infrastructure from unauthorized access, alteration, disabling and deletion. Separate log administration from monitored operational duties and preserve evidence of logging changes.
Why this matters
A privileged operator who can both change a model and erase its records can conceal misuse. Audit confidence depends on the infrastructure retaining trustworthy events.
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
Apply to log collectors, storage, export pipelines, keys, retention settings and monitoring consoles. Include provider-managed evidence paths.
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 security logging owner protects infrastructure; platform teams maintain availability; independent reviewers oversee sensitive changes.
Implementation guidance
- Identify audit infrastructure and privileged control paths.
- Restrict writes, deletion and configuration changes with separated roles.
- Protect integrity and retention using appropriate storage controls.
- Alert on logging interruption and test unauthorized modification.
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 |
|---|---|
| Operator attempts log deletion | Access is denied or independently controlled. |
| Logging configuration changes | Authorized event and alert are retained. |
| Collector unavailable | Visibility loss is detected and escalated. |
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
- Logging access matrix
- Retention configuration
- Integrity tests
- Change alerts
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
Preserve remaining evidence, isolate compromised administration and document the period and systems whose audit coverage is uncertain.
Example in practice
Hypothetical example: A serving administrator cannot remove security events from the separate log account. Disabling an exporter generates an alert through an independent health check.
Related controls
- CF12-S03: AI System Separation of Duties
- CF04-001: AI Audit & Assurance Policy
- CF12-S06: AI Privileged Access Role Segregation
Framework mappings
- CSA AICM (IAM-12)
- 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 Security & Abuse Resistance
Version 1.0