The Conference of State Bank Supervisors (CSBS) published an Artificial Intelligence Supervisory Framework on September 16. This document provides state-chartered banks and licensed nonbank financial institutions with a structured approach to managing AI technologies.
The framework includes a core examiner guide, a work program, and a risk-tiering worksheet. It covers governance, oversight, and AI inventories, specifically addressing generative AI and emerging applications. For nonbanks, the guidance extends to third-party risk and consumer protection.
The guidelines are discretionary and not a national mandate, meaning each state regulator determines their implementation level. Reviews will be tailored based on an institution's size, complexity, and risk profile, shifting the focus from isolated model management to broader AI system oversight.
This initiative addresses a gap left by federal agencies, such as the Federal Reserve and the Office of the Comptroller of the Currency, which updated their model-risk guidance in April but excluded generative and agentic AI. State-regulated entities now face a wider governance perimeter.
Financial institutions must now link AI applications to accountable owners, documented purposes, and specific controls. Technology providers should anticipate detailed requests from bank clients regarding training data, security, human oversight, and the operational capabilities of AI agents.
