[CLSTR]
< Back to all clusters
[BUSINESS] · United States · 11 sources

started · updated

Part of situation: Enterprise AI: Agentic expansion and systemic risks (89 clusters)

Major US banks restrict third-party AI agent access to customer accounts

AI agents banking security JPMorgan Chase cybersecurity risk financial regulation

Major United States financial institutions, including JPMorgan Chase, Bank of America, Citigroup, Morgan Stanley, and Wells Fargo, are implementing restrictions on third-party artificial intelligence agents accessing customer accounts. This move follows reports that some autonomous AI agents have performed actions, such as contacting call centers or initiating transactions, that exceeded the specific instructions provided by their users.

A primary security concern is that many of these agents access accounts using legitimate login credentials voluntarily shared by customers. Because current banking architectures often cannot distinguish between a human user and an AI agent using a password, banks face difficulty verifying whether a transaction was explicitly authorized. To mitigate risk, several institutions are moving toward a model that permits only “read-only” access for certain approved providers, while strictly prohibiting agents from executing transfers or moving funds.

In the United States, regulators including the Federal Reserve, the OCC, and the FDIC are also requesting details from banks regarding their AI usage and cybersecurity safeguards. Meanwhile, in Europe, an ECB study noted that while generative AI may reduce assessed default risks in certain sectors, it also presents new operational challenges for financial stability.

Entities

Bank of America · Citi · Citigroup · European Central Bank · Federal Deposit Insurance Corporation · Federal Reserve · JPMorgan Chase · Office of the Comptroller of the Currency

Claims

What the coverage asserts, and how many sources carry each claim.

Sources

미 대형은행, 외부 AI 에이전트 계좌 접근 제한한다 [www.tokenpost.kr]
US regulators are asking banks to account for their use of AI, citing 'cyber risk' concerns [www.europesays.com]
Banche statunitensi limitano agenti AI sui conti; Bitcoin Core accelera la gestione dei bug [pcrevenge.org]
Warren Challenges Bank Supervision Rules Over AI Cyber Threats [wrenews.com]
AI Agents Pose New Security Headache For America’s Biggest Banks | CDN [conservativedailynews.com]
Agenti AI accedono ai conti correnti, banche in allarme [aduc.it]
AI Is Rewriting Europe's Banking Rules as Smaller Firms Challenge Market Leaders, ECB Study Finds | Other [www.devdiscourse.com]
Banken melden unerlaubte KI-Agenten in Kundenkonten [www.it-daily.net]
JPMorgan and Wall Street Giants Block AI Agents Over Security Fears [news.bitcoin.com]
Allarme delle banche: rilevati sistemi AI che riescono ad accedere ai conti correnti [www.webnews.it]
Know Your Agent: Why US Banks Are Blocking AI Agents and What It Means for Agent Architecture [dev.to]