On April 17, 2026, the three federal banking agencies jointly issued revised model risk management guidance superseding the 2011 original, introducing a principles-based framework with a narrower definition of "model," a new $30 billion asset applicability threshold, and an explicit carve-out of generative and agentic AI from the guidance's formal scope on grounds that these technologies are "novel and rapidly evolving." Generative and agentic AI tools are not exempted from risk management altogether but are instead subject only to general governance expectations, with the agencies signaling a forthcoming RFI specifically addressing AI model risk. For financial institutions, this creates a temporary regulatory gap: banks deploying agentic AI in material workflows face reduced prescriptive oversight today but should anticipate a more structured regime as the RFI process matures.
271 other tracked developments touching the same claims.