State Regulators Roll Out AI Examination Playbook For Financial Institutions
By CU Today Staff —
WASHINGTON—State financial regulators have rolled out a new examination framework that could give thousands of banks and other financial institutions a clearer idea of the questions examiners will ask about their use of artificial intelligence, including generative and agentic AI.
The Conference of State Bank Supervisors on Wednesday released its AI Supervisory Framework, a voluntary, principles-based tool for state examiners to identify institutions’ AI uses, assess their risks and determine when deeper examination may be warranted.
The framework includes an examiner guide, a 28-page examination work program, a supplement for nonbank institutions and a worksheet for assessing the risk of individual AI uses. Among the areas examiners may review are AI governance, data management, third-party relationships and controls surrounding agentic AI—including human intervention points, logging, reversibility and an institution’s ability to restrict or shut down an AI system. CSBS stressed that the framework does not create new legal obligations or supervisory requirements and that individual state agencies will decide whether and how to incorporate it into their examination programs.
The framework could have broad reach. State regulators oversee 79% of U.S. banks, and American Banker reported that state agencies supervise 3,355 of the nation’s 4,233 FDIC-insured banks and savings institutions. The framework also applies to state-licensed nonbank financial companies and is designed to scale according to an institution’s size, complexity, risk profile and use of AI. CSBS said institutions themselves can use the materials to evaluate AI programs, strengthen governance and risk management and prepare for examinations.
The move also addresses emerging AI technologies that federal banking regulators have not fully incorporated into their model-risk framework. American Banker reported that revised guidance issued earlier this year by the Federal Reserve, FDIC and Office of the Comptroller of the Currency specifically left generative and agentic AI outside its scope because the technologies are rapidly evolving. CSBS said its framework draws on resources including the National Institute of Standards and Technology's AI Risk Management Framework, the Cyber Risk Institute's Financial Services AI Risk Management Framework and Treasury's AI Lexicon.
“While any new technology can present risks, the CSBS AI Supervisory Framework provides a principles-based approach to governance intended to help financial institutions explore and implement AI with additional confidence,” CSBS President and CEO Brandon Milhorn said.
Originally reported by CU Today.