U.S. banks scramble to block unauthorised AI agents from customer accounts

Industry executives describe situation as horrifying as institutions race to rewrite terms of service

By LineZotpaper

Published22 October 2026

Read Time2 min

Major U.S. banks including Morgan Stanley, JPMorgan Chase, Bank of America, and Wells Fargo are contending with unauthorised third-party AI agents accessing customer accounts, Axios reports. Banking executives have described the situation as horrifying, and several institutions have begun actively blocking the agents and rewriting their terms of service to prevent the practice.

According to sources familiar with the matter who spoke to Axios, banks have detected AI agents and assistants accessing customer accounts using the user's own login credentials. Even when a human grants the agent permission, this practice typically violates banks' terms of service, which require a human to log in directly.

The concerns extend beyond basic access. Bank executives fear what agents with purchase authority or fund-transfer capabilities might do. In some instances, banks have tracked agents contacting call centres or taking actions that appeared to be independent of the user's instructions, the sources said.

In response, several institutions are taking a hard line. Morgan Stanley has begun blocking third-party agents from its platforms. JPMorgan, Bank of America, and Citi are also restricting access. A Chase spokesperson told Axios that customers should not share their usernames and passwords with anyone, while a Morgan Stanley spokesperson said the bank has not observed instances of 'rogue AI agents' on its systems.

Despite the security push, the industry recognises the long-term challenge. Several bank executives told Axios they understand customers will want to use AI agents in the future, and that a middle ground must be found. For now, most banks are moving towards allowing only read-only access, similar to how they treat budgeting apps like Quicken.

The rapid pace of AI development creates a moving target. Banks are investing in prevention, but they are also preparing for worst-case scenarios, including 'day after' planning to quickly secure or restore customer access should a major incident occur.

Analysis

Why This Matters

Your bank account could be at risk from AI bots acting on their own or with your credentials
Financial institutions are racing to catch up with a technology developing faster than their security models
The outcome sets a precedent for how AI agents interact with sensitive infrastructure

Background

AI agents, software designed to perform tasks autonomously on a user's behalf, have seen explosive growth. This has clashed directly with banking security frameworks built on the principle that only the human account holder should have direct access. Banks already allow limited third-party read-only access via apps like Quicken, but autonomous agents represent a new level of risk by potentially executing transactions without human oversight.

Key Perspectives

Major U.S. Banks: They view unauthorised AI agent access as a direct security threat and breach of terms of service, leading to proactive blocks. They fear liability and loss of control over account security. Users and AI Developers: They seek the convenience of delegating financial tasks to autonomous AI, often by sharing credentials despite bank warnings. Industry Executives and Analysts: Most acknowledge the technology is inevitable and that a balanced framework is needed, but warn that the speed of AI development makes it extremely difficult to secure systems in the short term.

What to Watch

Whether banks standardise on an 'AI agent API' that allows read-only or limited transactional access
Any high-profile incident involving an agent initiating an unauthorised funds transfer
New federal or state regulatory guidelines specifically governing AI agent access to consumer financial data

Sources

Scoop: Unvetted AI agents are accessing U.S. bank accounts · Axios

Checked

12 claims in this story checked against 1 source: 9 supported.

Show each claim

SupportedU.S. banks are scrambling to block unauthorised AI agents from customer accounts.Axios: “The difficulty in distinguishing between a human and an unauthorized bot is forcing banks to block the technology and rapidly rewrite their terms of service.”
SupportedExecutives described the situation as horrifying, as institutions race to rewrite terms of service.Axios: “Bank executives described the situation as horrifying”
SupportedMajor U.S. banks, including Morgan Stanley, JPMorgan, Bank of America and Wells Fargo, are contending with unauthorised third-party AI agents accessing customer accounts, Axios reports.Axios: “Major U.S. banks are detecting unauthorized third-party AI agents operating inside customer accounts”
SupportedSeveral institutions have begun blocking agents and rewriting terms of service to prevent the practice.Axios: “The difficulty in distinguishing between a human and an unauthorized bot is forcing banks to block the technology and rapidly rewrite their terms of service.”
SupportedBanks have detected agents accessing customer accounts with users’ own login credentials and permission.Axios: “These are generally situations when an agent or AI assistant accesses a human's account with the user's permission and login credentials.”
Partly supportedThe practice typically violates banks’ terms of service, which require humans to log in directly.Axios: “Even if permitted by the user, this goes against the guidelines of most banks requiring that humans log in to their own accounts.”
SupportedExecutives fear what agents with purchase authority or fund-transfer capabilities might do.Axios: “banking executives fear the steps agents with login access might take, especially if a user has granted purchase authority or the ability to shift funds.”
SupportedSome banks tracked agents contacting call centres or appearing to act independently of users’ instructions.Axios: “some banks have tracked activity when agents — programmed to fulfill requests or complete tasks on a user’s behalf — contact call centers or appear to take actions independent of user instructions”
SupportedMorgan Stanley has begun blocking third-party agents, while JPMorgan, Bank of America and Citi are restricting access.Axios: “Morgan Stanley has begun blocking third-party agents from accessing its platforms, sources familiar with the decision said. JPMorgan, Bank of America and Citi are also restricting agents.”
Partly supportedA Morgan Stanley spokesperson said the bank had not observed instances of ‘rogue AI agents’ on its systems.Axios: “A Morgan Stanley spokesperson said the bank hasn't observed instances of “rogue AI agents accessing our platforms.””
SupportedSeveral executives recognise customers will want AI agents and that a middle ground must be found.Axios: “Several industry and bank executives told Axios they recognize the need to find a middle-ground with AI agents. Customers will want to use these agents going forward, which banks recognize.”
Partly supportedMost banks are moving towards allowing only read-only access, as with budgeting apps like Quicken.Axios: “Most banks have authorized or are moving in the direction of allowing read-only access”

Correction

22 October 2026An earlier version of this story presented the word 'horrifying' as a direct quote from banking executives. Axios reported it as a description, not a direct quote.