US bank regulators examine controls over autonomous AI agents

Three US federal banking regulators are asking lenders how they monitor, audit and control decisions made by autonomous artificial intelligence (AI) agents, including how they would shut down these systems, according to a Reuters report dated October 13.

The Office of the Comptroller of the Currency (OCC), the Federal Deposit Insurance Corporation (FDIC) and the Federal Reserve are seeking information from banks on AI oversight, decision-making authority, audit trails, human intervention and emergency shutdown procedures. The inquiries also cover the use of Chinese-developed AI technology and AI agents in cybersecurity.

Regulators have described the exercise as information gathering rather than the introduction of new requirements.

The move comes as banks face challenges involving legacy technology, limited in-house AI expertise and data governance. In April 2026, the three agencies issued revised model risk management guidance, excluding generative and agentic AI from its scope because of their rapidly evolving nature.

The agencies have indicated that broader risk management and governance practices should guide the adoption of these technologies.

The inquiries highlight the growing importance of board-level oversight in ensuring AI accountability, effective human supervision, robust audit trails and controls to prevent autonomous systems from operating beyond their authorised limits.