Federal regulators, including the Federal Reserve, the Office of the Comptroller of the Currency, and the FDIC, are actively questioning banks on their deployment of agentic AI. These systems, designed to operate with minimal human intervention, have triggered alarms regarding the ability of lenders to audit decisions, maintain "kill switches," and prevent unauthorized actions. Officials are specifically probing whether banks utilize Chinese-developed technology or rely on autonomous agents for critical cyber defense, fearing these tools could be weaponized or escape oversight.

Across the Atlantic, the Bank of England and the European Central Bank are demanding that lenders sharpen their disaster recovery strategies. European supervisors are pushing for "bare metal" recovery plans—independent, isolated systems capable of functioning even if a bank’s primary infrastructure is compromised. The urgency follows recent incidents, including hacks on South Korean banks and the platform Hugging Face, which underscored the potential for rogue AI to breach internal security. As banks scramble to adopt these models to stay competitive, experts warn that the speed of implementation is currently outpacing the development of necessary safety controls.