AI Control Library / CF-12: Security & Abuse Resistance
Designation: Core | Control types represented: Operational
Control requirement
Establish and maintain an identity and access management policy covering people, services, agents and suppliers that interact with AI systems. Define approval, privilege, authentication, review and revocation requirements.
Why this matters
AI access extends beyond the chat interface to data stores, model registries, tools and execution infrastructure. Unclear responsibilities allow permissions to accumulate across those paths.
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
Apply across development and production, including provider consoles, machine identities and customer-support access. Translate applicable external obligations into specific policy requirements where relevant.
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 policy owner maintains requirements; AI service owners implement them; identity teams operate shared controls.
Implementation guidance
- Inventory AI access domains and accountable system owners.
- Define lifecycle rules for human, workload and agent identities.
- Specify approval, review, exception and emergency-access processes.
- Publish responsibilities and assess actual implementation against policy.
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 |
|---|---|
| System sample | All access domains have assigned policy requirements. |
| Exception sample | Approval, justification and expiry are recorded. |
| Access review | Findings produce tracked corrective action. |
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
- Approved policy
- Scope inventory
- Exception records
- Assessment findings
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
Assign an owner to missing requirements, restrict unjustified access and prioritize gaps affecting sensitive data or consequential actions.
Example in practice
Hypothetical example: An organization discovers that its AI policy covers employee logins but omits retrieval service accounts. The revised policy assigns owners and review duties for those identities.
Related controls
- CF12-002: AI System Least Privilege Access Control
- CF12-S02: AI System Identity Inventory
- CF12-S05: AI System Access Revocation
Framework mappings
- CSA AICM (IAM-01)
- NIST AI RMF (GOVERN 1.1)
- ISO 42001 (Cl. 8.2)
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 Security & Abuse Resistance
Version 1.0