From Decrypt · By Jose Antonio Lanz
In brief
The OCC has issued its first framework for examiners assessing banks' use of AI, focusing on consequences and autonomy rather than model sophistication.
The guidance explicitly addresses outside AI agents, echoing concerns raised by major banks.
The framework is nonbinding, but provides examples of safeguards for autonomous payments, third-party access, and human oversight.
The U.S. Office of the Comptroller of the Currency issued its first supervisory framework for artificial intelligence in banking today, telling examiners to focus less on how powerful an AI system is and more on what happens when it makes a mistake.
The OCC's framework addresses that very concern, although the regulator has not linked its issuance to those specific incidents.
In a section covering operational resilience, the agency warns that external agents operated by customers, fintechs, or other businesses could create risks for banks even when the banks don't control those systems.
PredictEconomyGold Highs in November
4,10086%YesNo4,20052%YesNo4,30032%YesNo4,40017%YesNo4,000-%YesNo+3 more outcomes+2 more outcomesNEW$200.00 volumeCloses: Dec 1, 2026, 04:59 UTCCloses: Dec 1, 2026, 04:59 UTCIt suggests examiners consider measures including identity verification, delegated permissions, transaction limits, monitoring, and the ability to suspend or revoke access.
The guidelines aim to tackle the dilemma of how to let customers use AI assistants to manage their finances without giving autonomous software unrestricted access to accounts that banks are already facing.
The new OCC guidance arrives less than two weeks after reports that major banks were discovering unauthorized third-party AI agents operating inside customer accounts, sometimes taking actions that appeared to go beyond users' instructions.
Morgan Stanley was blocking outside agents, while JPMorgan, Bank of America, and Citi were restricting access, according to an Axios report on October 22. Bank executives described the situation as "horrifying," raising questions about who controls an autonomous agent once it has access to someone's money.
AI that moves money is riskier than one that writes emails
The OCC's framework makes a distinction between an AI system's technical sophistication and the damage it could cause.
A simple agent authorized to initiate financial transactions may present greater risk than a more advanced model limited to writing internal documents, the regulator said.
Under the 22-page framework, examiners can consider the consequences of failure, whether an action can be reversed, and how much control humans retain over the system.
The agency outlines four illustrative risk tiers. Fully autonomous payment agents and real-time credit decision systems fall into the highest category, while basic internal productivity tools occupy the lowest.