CiberLATAMbywhalemate

OCC extends deadline, sets AI oversight framework

The OCC and other U.S. regulators extended the comment period on third-party risk guidance to Dec. 16, 2026, and issued an AI exam framework.

Whalemate Labs · AI-assisted researchPublished:2 min read

The OCC, along with the FDIC, the Federal Reserve and the NCUA, extended the deadline to submit comments on a proposed third-party risk management guide for U.S. financial institutions until Dec. 16, 2026. The office also released a supervisory framework for examiners to assess banks’ use of generative and agentic artificial intelligence.

The OCC, together with the FDIC, the Federal Reserve and the NCUA, has extended to Dec. 16, 2026, the deadline for public comments on proposed guidance for managing risks tied to third-party relationships at U.S. financial institutions. At the same time, the OCC released a supervisory framework for examiners to assess banks’ use of generative and agentic artificial intelligence, with a focus on national banks and other entities under its authority.

What changed in the comment period?

The deadline to weigh in on the proposed guidance moved from Nov. 16 to Dec. 16, 2026, according to the FDIC and the OCC. The extension adds 30 days and applies jointly to the four agencies involved in the initiative.

The OCC announced the extension in a notice dated Nov. 13, 2026. The FDIC, for its part, issued a separate notice confirming the same schedule change for the third-party risk management guidance.

What does the OCC framework for AI include?

The framework is designed to help OCC examiners assess banks’ use of generative and agentic AI, and the agency said it is a supervisory resource, not a new enforceable regulatory requirement.

According to Arnold & Porter, the scope covers credit decisions, payments, fraud detection, customer service and risk management. Reed Smith added that it also covers operations, market activity and other day-to-day banking functions.

Arnold & Porter also said the framework is aimed at national banks, federal savings associations and branches and agencies of foreign banks. The firm noted that the interagency model risk guidance issued in April 2026 excluded generative and agentic models because they are novel and evolving quickly, while non-generative and non-agentic models remain covered by that guidance.

What regulatory scope does the document have?

The OCC said the framework does not create new enforceable regulatory requirements. Arnold & Porter also described it as nonbinding, with no new supervisory expectations, enforceable standards or prescriptive requirements, and said it does not expect banks to use it to self-assess their AI risk management.

That leaves the document as a supervisory tool for examiners, while the third-party risk discussion remains open until mid-December. The two moves show a simultaneous shift in the U.S. banking prudential agenda, with attention on outside vendors and the use of generative and agentic AI inside financial institutions.

Sources

View all