AI Control Library / CF-03: AI Supply Chain & Third-Party Risk
Designation: Core | Control types represented: Operational
Control requirement
Assess AI supply chain risks before material integration and after relevant changes, covering supplier security, provenance, data handling, model behavior, service continuity, concentration and customer control limitations.
Why this matters
A technically strong model may still be unsuitable because the organization cannot protect its data, control updates or recover from loss of the service. The assessment must consider the complete dependency and its actual use.
How the control types work together
Operational controls establish the supplier assessment, ownership, agreement and review processes described here. They work when the assigned parties perform the required activities and use evidence to guide decisions.
When it applies
New and changing AI components and providers, including open-weight downloads, embedded features, service intermediaries and critical upstream dependencies.
Tailor review depth to information sensitivity, granted authority and business dependence. Record what the organization can test directly, what relies on provider evidence and how material visibility gaps affect the acceptance decision.
Ownership
The internal service owner is accountable; supplier-risk, security, privacy and domain specialists assess evidence and treatment options.
Implementation guidance
- Define the deployment role, information exposure, permissions and criticality; identify the consequences of supplier error, compromise or withdrawal.
- Evaluate relevant evidence on security practices, model provenance, data terms, change control, resilience and incident cooperation.
- Assess gaps, uncertainty, common dependencies and substitutability; identify controls available at the customer’s own boundary.
- Record approval conditions, treatment owners and reassessment triggers, checking that accepted limitations do not contradict the intended use.
How to verify it
Use representative deployment records and authorized controlled tests. Establish the assessed service, artifact version and acceptance conditions before interpreting results; do not extend evidence beyond the boundary it covers.
| Test | Expected result |
|---|---|
| Trace a high-impact supplier approval | Deployment-specific risks and evidence support the decision. |
| Inspect missing provider evidence | The limitation has a treatment or restriction rather than an assumed favorable result. |
| Simulate loss or compromise of a critical dependency | A feasible containment, fallback or suspension decision is defined. |
Record a finding when a required result cannot be demonstrated or actual practice differs materially from the accepted conditions. Assign corrective action and verify the affected boundary before closing the issue.
Evidence to retain
- Supply chain risk assessment
- Supplier and technical evidence
- Treatment and approval decisions
- Reassessment and dependency analysis
Protect supplier-confidential information, credentials and personal data in these records. Preserve enough version and scope information to establish which service or artifact the evidence supports.
If the control fails
Escalate material uncertainty or risk outside tolerance to the business risk authority. Narrow, delay or reject the integration where available controls cannot support the intended exposure.
Example in practice
Hypothetical example: A provider offers strong model quality but cannot prevent customer prompts from being retained under the proposed service terms. The buyer restricts the use to nonsensitive inputs rather than approving confidential-document processing.
Related controls
- CF03-005: AI Supply Chain Inventory
- CF03-007: AI Supply Chain Contractual Requirements
- CF03-012: AI Supply Chain Data Security Assessment
Framework mappings
- CSA AICM (STA-09)
- NIST AI RMF (MAP 1.6, MEASURE 2.7)
- EU AI Act (Art. 9(2)(b))
- ISO 42001 (Cl. 6.1, A9.1)
Review and maintenance
Review at the interval approved by the control owner and after relevant supplier, component, contract, access or deployment changes. Reassess material incidents and new evidence, updating acceptance conditions and verifying corrective actions where required.
Back to AI Supply Chain & Third-Party Risk
Version 1.0