Fed and OCC Issue Model-Risk Guidance Excluding Generative and Agentic AI

For the past 15 years, U.S. banks have primarily relied on the Federal Reserve’s 2011 SR 11-7 guidance to manage risks in models used for credit, pricing and stress testing. As generative and agentic AI move into lending, fraud prevention and customer service, hallucinations, data leaks and questions of accountability are also increasing litigation and governance pressures. The scope of official guidance is therefore critical to bank compliance and consumer protection.

On April 17, 2026, the Federal Reserve, Office of the Comptroller of the Currency (OCC) and Federal Deposit Insurance Corporation (FDIC) jointly issued SR 26-2, replacing the 2011 SR 11-7 and 2021 SR 21-8 guidance and focusing on banks with more than $30 billion in total assets. The new guidance covers traditional and non-generative AI models but excludes generative and agentic AI because the technologies are novel and evolving rapidly. For now, banks can determine the relevant controls themselves.