FSA AI / Model Governance Expectations for Financial Institutions · vFSA AI Discussion Paper (Ju…
FSA AI / Model Governance Expectations for Financial Institutions
FSA AI / Model Governance Expectations for Financial Institutions — 90% of in-scope requirements covered.
5 requirements · 5 in scope (4 enforced · 1 designed). The 90% is weighted over the in-scope base.
Source: Japan Financial Services Agency (FSA) AI Discussion Paper (June 2024) and the model-risk / system-risk-management governance expectations in the FSA's Comprehensive Guidelines for Supervision of financial institutions. KYE™ maps the AI governance, model-risk-management, human-oversight, explainability, and operational-resilience expectations that bind an AI-supported financial action onto runtime evidence.
By category
| Category | Reqs | Enforced | Designed | Advisory | Deferred | Coverage |
|---|---|---|---|---|---|---|
| AI governance & accountability | 1 | 1 | 0 | 0 | 0 | 100% |
| Model risk management | 1 | 1 | 0 | 0 | 0 | 100% |
| Human oversight | 1 | 1 | 0 | 0 | 0 | 100% |
| Explainability & customer disclosure | 1 | 1 | 0 | 0 | 0 | 100% |
| Operational resilience & incident management | 1 | 0 | 1 | 0 | 0 | 50% |
Every requirement → the KYE™ artefact that enforces it
| ID | Title | Status | KYE™ enforcement |
|---|---|---|---|
fsa-guidelines-ai.governance-accountability |
FSA AI expectations — establish an AI governance framework with clear accountability for the development, deployment and oversight of AI in financial services | enforced | audit_events: kye.compliance.attestation.v1, kye.purpose.request.v1engines: internal, internalconstitution_refs: constitution/12-PURPOSE-PERMISSION.md, constitution/00-INDEX.md |
fsa-guidelines-ai.model-risk-management |
FSA AI expectations — manage model risk across the AI lifecycle, including validation, performance monitoring and data-quality controls | enforced | audit_events: kye.risk_assessment.v1, kye.resilience.drift.detected.v1engines: internal, internal, internalconstitution_refs: constitution/13-RESILIENCE-LOOP.md |
fsa-guidelines-ai.human-oversight |
FSA AI expectations — maintain meaningful human oversight and the ability to intervene in AI-driven decisions affecting customers | enforced | audit_events: kye.purpose.admissibility.v1, kye.evidence.decision_map.v1engines: internalconstitution_refs: constitution/36-GOVERNEDUI.md, constitution/12-PURPOSE-PERMISSION.md |
fsa-guidelines-ai.explainability-disclosure |
FSA AI expectations — ensure AI-driven decisions affecting customers are explainable and that customers are appropriately informed | enforced | audit_events: kye.evidence.decision_map.v1, kye.evidence.pack.v1engines: internal, internalconstitution_refs: constitution/13-RESILIENCE-LOOP.md |
fsa-guidelines-ai.operational-resilience-incident |
FSA AI expectations — manage operational resilience and report material system / AI incidents to the FSA in a timely manner | designed | audit_events: kye.signal.incident.opened.v1, kye.compliance.attestation.v1engines: internal, internalconstitution_refs: constitution/13-RESILIENCE-LOOP.md, constitution/51-NO-SPOF.md |