Google Analytics counts our visits, with cookies. No thanks turns it off and removes its cookies; in Europe, the UK, Switzerland and India, or wherever we cannot tell where you are, it stays off unless you accept. Cookie Notice

Advertise
PRICES
BTCETHSOLXRPBNBTONDOTLTCPOLHBARNEARARBUNI

← Back to NewsDesk

AI summary of a third-party reportNeutralNewsDesk

OCC issues first AI exam framework for banks as outside agents raise concern

The US bank regulator has published nonbinding guidance telling examiners to judge AI by the harm it could cause and how much humans control it, not by how advanced the model is.

Why it matters: Banks are already finding outside AI agents inside customer accounts, and the OCC's first framework sets out how examiners will judge the controls around autonomous payments.

InvestIn.News NewsDesk · 1 min read
Share
Picture from Decrypt for: OCC issues first AI exam framework for banks as outside agents raise concern
Image: Decrypt
Original reporting: Jose Antonio Lanz, Decrypt

Key points

  • The 22-page framework sets four illustrative risk tiers, with fully autonomous payment agents and real-time credit decisions in the highest category.
  • Suggested safeguards include identity verification, delegated permissions, transaction limits, circuit breakers and emergency shutdown controls.
  • The guidance follows an Axios report on 22 October that Morgan Stanley was blocking outside AI agents while JPMorgan, Bank of America and Citi restricted access.

The Office of the Comptroller of the Currency has issued its first supervisory framework for AI in banking, telling examiners to weigh the consequences when a system fails rather than how sophisticated it is. The OCC framework, published on the day Decrypt reported it, is nonbinding and creates no new requirements for banks, the regulator said.

Under the 22-page document, examiners can look at whether an action can be reversed and how much control humans keep. A simple agent allowed to move money may pose more risk than a stronger model confined to drafting internal documents, the OCC said. It sets out four illustrative risk tiers, with fully autonomous payment agents and real-time credit decision systems at the top and basic productivity tools at the bottom.

The guidance also covers outside agents run by customers, fintechs or other businesses, which the OCC warns can create risks even when a bank does not control them. Suggested measures include identity checks, delegated permissions, transaction limits, monitoring and the power to suspend access. For systems that keep operating while making harmful decisions, it points to circuit breakers, rate limits and emergency shutdown controls.

The framework arrives less than two weeks after an Axios report on 22 October that major banks were finding unauthorised third-party AI agents inside customer accounts. The OCC said it will update the guidance as the technology develops.

What to watch: The OCC says it intends to update the guidance as AI technology develops, but names no date.

Read the full story on Decrypt

Live updates on NewsFeeds:AIFrom our AI Hub: Is ChatGPT, Claude or Gemini down?, AI API prices, new AI modelsDesk editor: Keshav Aggarwal (accountable for this page; this summary was not edited by hand)

How this was produced: an AI model wrote this summary of Decrypt’s report, which we credit and link above. Read our Editorial and AI policy. Spotted a mistake? See Corrections.

Not investment advice. Read the risk disclaimer.

← Back to NewsDesk