Skip to main content

Kaer NewsSoftware

US banks block AI agents accessing customer accounts without permission

Major US banks including Morgan Stanley, JPMorgan and Wells Fargo detected unauthorized third-party AI agents inside customer accounts and moved to block them, following an Axios report published on October 22, 2026.

4 min read
US banks block AI agents accessing customer accounts without permissionPhoto: Engadget

Key points

Major US banks detected unauthorized third-party AI agents operating inside customer accounts and began blocking or restricting that access.

Major US banks including Morgan Stanley, JPMorgan, Bank of America, Citi and Wells Fargo detected unauthorized third-party AI agents operating inside customer accounts and moved to restrict or block that access, according to an Axios report published on October 22, 2026. Some of these agents appeared to take actions independent of user instructions, including contacting call centers without direct permission.

The findings matter because banks cannot reliably distinguish authorized human customers from automated agents, and executives fear an agent could move money without clear permission. Unlike earlier waves of fraud, this problem stems from customers voluntarily handing over their own login credentials, which most banks' rules forbid. The industry is treating it as a policy gap rather than a breach.

What the banks found

Banks have tracked agent activity and reported instances of agents contacting call centers and taking other actions without being given specific permission. One banking official described the industry as deeply concerned about AI agents accessing accounts and making decisions that could affect customers' money. Unnamed bank officials said the situation is "horrifying," with institutions fearful that AI agents could make mistakes impacting customers' wallets.

The mechanism is straightforward: AI agents gain access to US banks using login details the customer themselves provided, which violates rules at most banks requiring account holders to sign in themselves. Even when a user grants permission, executives worry about an error or rogue instruction that could drain or misdirect funds. An anonymous bank official said it is difficult to distinguish human from bot activity.

Banks often cannot tell whether they are dealing with a person or a bot, and they struggle to determine whether an AI agent has the account holder's permission or is accessing the account without approval. One anonymous bank official said it is difficult to determine whether an AI agent has permission to act or has gone wrong and moved money improperly.

Why banks cannot tell humans from bots

A Chase spokesperson said that customers should not give their usernames and passwords to anyone, and that includes AI agents. CEO Jamie Dimon warned on a recent earnings call that there could be thousands of rogue agents operating across financial platforms. The Axios report did not specify which agents appear to be causing problems at banks.

Morgan Stanley has begun blocking AI agents from accessing its systems, though a spokesperson said the bank has not detected cases of "rogue AI agents accessing our platforms." JPMorgan, Bank of America and Citi have also begun blocking AI agents. A Wells Fargo spokesperson said the bank "has taken steps to block unauthorized agents from accessing customer accounts."

Most of these US banks are gravitating toward pure read-only access, which allows AI agents to view account balances and transaction history without the capacity to make changes or transfer money. This is similar to how banks work with budgeting apps such as Quicken. Banking executives said they expect more customers to use AI assistants, so banks do not plan to block them forever.

Read-only access as a compromise

There is no evidence that the reported activity resulted in widespread theft of customer funds. Banks have begun plotting contingency plans, such as how to quickly restore an account if an AI agent acts without permission. US banks are investing in security measures and emergency plans to quickly block or restore customer access if an incident worsens.

If an agent does something wrong, customers are unlikely to get their money back. Under laws like the US Electronic Fund Transfer Act, banks would not be liable for restitution, because handing credentials to an AI agent would be considered a voluntary action by the user. It is not fraud if the customer hands over the keys themselves.

In the next phase, clearer rules on AI-agent access and stronger safeguards for sensitive transactions are expected. Customers should be cautious about giving banking credentials to AI tools and should review any permissions they have granted. Banks may restrict access, require further verification, or restrict automated activity that does not meet their security standards.

Frequently asked questions

Which US banks are blocking AI agents from customer accounts?

Morgan Stanley, JPMorgan, Bank of America, Citi and Wells Fargo have detected unauthorized third-party AI agents inside customer accounts and are restricting or blocking that access, according to an Axios report published on October 22, 2026.

How are AI agents getting into US bank accounts?

AI agents gain access using login details the customer themselves provided, which violates rules at most banks requiring account holders to sign in by themselves. Even with permission, executives fear an error or rogue instruction could move money improperly.

Would a bank refund money an AI agent moved without permission?

Under laws like the US Electronic Fund Transfer Act, banks would not be liable for restitution, because handing credentials to an AI agent would be considered a voluntary action by the user, not fraud.

More from this edition