The Office of the Comptroller of the Currency today issued guidance for evaluating banks’ use of artificial intelligence, directing examiners to focus on the risks posed by applications rather than treating the technology as a separate category of financial risk.
The guidance covers national banks, federal savings associations and federal branches of foreign banks. It addresses AI and autonomous agents in financial services, including systems that can make decisions or initiate transactions.
Examiners are directed to consider an application’s potential consequences, autonomy and human oversight. Even relatively simple systems could pose significant risks if authorized to perform sensitive financial activities.
An AI agent able to initiate transactions, for example, could warrant more scrutiny than a sophisticated model used for internal research. Higher-risk uses may call for tighter controls and more human supervision.
The guidance does not create new legally enforceable requirements. It supplements existing supervisory practices concerning governance, technology, compliance and operational risk.
The framework comes as banks increasingly explore autonomous AI tools while regulators assess the potential for cybersecurity breaches, operational failures and erroneous transactions.