Central Banks Ask Banks: How Do You Stop Agentic AI?

Reuters reported, citing six people familiar with the matter, that banking regulators in the United States and Europe have intensified their questioning of banks regarding the use of advanced and agentic artificial intelligence.

The Federal Reserve, the Office of the Comptroller of the Currency, and the Federal Deposit Insurance Corporation are all asking banks about the limits they set for autonomous systems, their auditability, and the ability to shut them down.

A source stated that the Bank of England and the European Central Bank are pressing lenders on recovery plans, including "bare-metal recovery," with an ECB deadline of late October for plans to address AI-related cybersecurity risks.

According to Reuters, this move comes at a fact-finding stage that does not yet impose binding rules, amid growing concern over the stability of the financial system as reliance on AI expands.

What do these terms mean?

Agentic AI: AI systems that execute tasks and take actions on their own without direct human intervention at every step.

Auditability: The ability to trace a system's decisions, review its steps, and understand the reasons behind its actions.

Bare-metal recovery: Rebuilding systems from scratch on bare hardware following a complete outage, serving as a contingency plan for banks.