AI Control Library / CF-05: Data Governance & Privacy
Designation: Core | Control types represented: Operational and Technical
Control requirement
Maintain an inventory of data used and created by AI, identifying sources, owners, sensitivity, purposes, locations, recipients, retention and relevant derivatives. Reconcile the inventory with deployed workflows and material changes.
Why this matters
Knowing the source database is insufficient when an AI workflow also creates vector indexes, caches, debugging records and supplier copies. Untracked information cannot be reliably protected, located or removed.
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
Training, adaptation, retrieval, inference, evaluation and monitoring data, including prompts, outputs, embeddings and backups 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 data owner is accountable; stewards and system teams maintain records and reconcile technical evidence.
Implementation guidance
- Record material datasets and data stores with ownership, classification, purpose and deployment references.
- Include generated and intermediate copies, hosted-service retention and relevant third-party destinations.
- Reconcile inventories using authorized discovery, configuration review and representative data-flow tracing, documenting visibility gaps.
- Update after changes and retirement, linking records to retention, rights handling and disposal responsibilities.
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 |
|---|---|
| Trace a representative item through the workflow | Material stores, derivatives and recipients appear in the inventory. |
| Inspect a new cache or index | Ownership and handling requirements are recorded. |
| Locate data for an authorized review | The inventory supports retrieval without relying solely on an engineer’s memory. |
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
- Data inventory and ownership records
- Deployment and store references
- Reconciliation results
- Change and retirement history
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 unknown sensitive-data stores or recipients to data and security owners. Contain unapproved processing and extend the inventory before relying on it for assurance or deletion claims.
Example in practice
Hypothetical example: A document assistant’s source files are inventoried, but an evaluation service retains prompt samples. Reconciliation adds that store and assigns retention and access controls.
Related controls
- CF05-003: AI Data Classification
- CF05-004: AI Data Flow Documentation
- CF05-011: AI Data Retention & Deletion
Framework mappings
- CSA AICM (DSP-03)
- NIST AI RMF (MAP 1.1)
- EU AI Act (Art. 10(2))
- ISO 42001 (Cl. 8.2, A7.1)
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 Data Governance & Privacy
Version 1.0