AI Control Library / CF-07: Human Oversight & Escalation
Designation: Supplemental | Control types represented: Operational and Technical
Control requirement
Adjust or revoke AI access promptly when personnel roles or authorization change, covering accounts, sessions, tokens, credentials, delegated rights and owned automations. Transfer legitimate operational responsibilities explicitly.
Why this matters
Removing a person from a group may leave active sessions, API tokens or scheduled agents working with their authority. Role changes need to address continuing access as well as the account’s visible membership.
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
Employee, contractor and supplier role changes, departures and emergency access termination affecting AI systems or supporting 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
The identity owner executes changes with HR and managers; service owners reconcile automations and shared or delegated access.
Implementation guidance
- Receive authoritative role-change information and determine required timing according to risk and organizational procedures.
- Inventory affected identities, sessions, tokens, tool grants and delegated automations, including external AI services.
- Revoke obsolete authority and rotate exposed shared secrets where needed, transferring approved workloads to controlled ownership.
- Test effective denial and reconcile remaining legitimate access, recording exceptions and incomplete coverage.
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 |
|---|---|
| Access a protected AI service after revocation | Obsolete accounts or sessions cannot continue the revoked activity. |
| Inspect scheduled and delegated work | It is stopped or transferred under approved authority. |
| Review a partial role change | Unnecessary rights are removed without granting broader replacement access. |
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
- Role-change notification
- Access and delegation inventory
- Revocation and transfer records
- Denial and exception 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
Escalate persistent access to security and identity authorities. Apply independent service or network restrictions while resolving incomplete revocation.
Example in practice
Hypothetical example: A departing operator’s login is disabled, but an API token still controls an agent. Reconciliation revokes the token and transfers the required scheduled task to an accountable service identity.
Related controls
- CF12-002: AI System Least Privilege Access Control
- CF07-S05: AI Personnel Asset Return on Departure
- CF07-S09: AI Personnel Roles & Responsibilities Documentation
Framework mappings
- CSA AICM (HRS-06)
- NIST AI RMF (GOVERN 5.2)
- ISO 42001 (Cl. 7.2)
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 Human Oversight & Escalation
Version 1.0