On March 31, 2021 the Federal Reserve, OCC, FDIC, CFPB, and NCUA jointly published a request for information on how financial institutions use artificial intelligence and machine learning and how existing rules and guidance apply. It asked 17 questions covering explainability, data quality and bias, overfitting, cybersecurity, dynamic updating of models, third-party AI, community-bank use, and fair lending. The comment period, originally closing June 1, 2021, was extended to July 1, 2021. No rule followed, but it remains the clearest statement of the issues US bank regulators examine when reviewing AI.
| Document | 2021 Interagency AI RFI — Request for Information and Comment on Financial Institutions' Use of Artificial Intelligence, Including Machine Learning |
| Issued by | Board of Governors of the Federal Reserve System |
| Type | Consultation |
| Status | Final |
| Published | Mar 31, 2021 |
| Comment deadline | Jul 1, 2021 |
| Applies to | Banks, savings associations, credit unions, and other financial institutions supervised by the Fed, OCC, FDIC, CFPB, and NCUA (information-gathering only; no obligations) |
| Also issued as | 2021 Interagency AI RFI (OCC Bulletin 2021-17), FDIC FIL-20-2021, 2021 Interagency AI RFI |
| Official source | federalregister.gov ↗ |
| Use cases | Model risk management · Credit scoring & underwriting · Fair lending & discrimination · Fraud detection · Third-party & vendor AI · AI governance (general) |
What are the key points of 2021 Interagency AI RFI?
- Five agencies: Federal Reserve Board, OCC, FDIC, CFPB, and NCUA — the first joint federal AI consultation for financial institutions
- Catalogs AI use cases the agencies had observed: fraud identification, personalization of services, credit decisions, risk management, textual analysis, and cybersecurity
- Identifies explainability as a central risk, including the ability to give ECOA/Regulation B adverse-action reasons from complex models
- Flags data-related risks: quality, representativeness, and the potential for bias in alternative data and training sets
- Raises overfitting, dynamic updating (models that change as they ingest new data), and the challenge of validating vendor-supplied AI
- Asks specifically about community institutions' reliance on third-party AI providers and about fair-lending compliance
- Points to existing frameworks that already apply — SR 11-7 / OCC 2011-12 model risk guidance, third-party risk guidance, ECOA, FCRA, and UDAAP
What did 2021 Interagency AI RFI change for banks?
The RFI did not create new obligations, but it established the agencies' shared vocabulary for AI risk — explainability, data bias, dynamic updating, third-party dependence — that later surfaced in CFPB adverse-action circulars, the 2023 third-party risk guidance, and the 2026 model risk revision. Banks still cite it as the map of what examiners expect them to have answers for.
Did the 2021 interagency AI RFI result in a rule?
No. It was an information-gathering exercise by the Fed, OCC, FDIC, CFPB, and NCUA. The agencies used the responses to inform later actions, including the 2023 third-party risk guidance and the April 2026 revised model risk management guidance.
What AI risks did the 2021 RFI focus on?
Explainability, data quality and bias, overfitting, cybersecurity, dynamic model updating, oversight of third-party AI, community-bank reliance on vendors, and fair lending under ECOA and FCRA.
| Date | Document | Status |
|---|---|---|
| May 27, 2026 | Cook: Opportunities and Risks of AI (May 2026) — The Opportunities and Risks AI Presents for the Economy and Financial System — Governor Lisa D. Cook | Final |
| May 1, 2026 | Bowman: AI in the Financial System (May 2026) — Artificial Intelligence in the Financial System — Vice Chair for Supervision Michelle W. Bowman | Final |
| Apr 17, 2026 | SR 26-2 — Revised Guidance on Model Risk Management | In force |
| Jun 7, 2023 | SR 23-4 — Interagency Guidance on Third-Party Relationships: Risk Management | In force |
| Apr 9, 2021 | 2021 BSA/AML Model Risk Statement — Interagency Statement on Model Risk Management for Bank Systems Supporting BSA/AML Compliance | Superseded |
| Apr 4, 2011 | SR 11-7 — Supervisory Guidance on Model Risk Management | Superseded |
Follow every document these regulators publish
6 curated AI stories for banking executives · Every morning · Free
Subscribe to BankingNewsAI →