Is AI Taking Over Your Bank? Big Banks Just Slammed the Door

What happens when you hand an AI assistant the keys to your financial life and your bank cannot tell whether the software is following your instructions? That is the problem facing major U.S. banks, which are blocking outside AI agents from logging into customer accounts.

No report describes a customer locked out of an account, and none says an agent has drained one. In the cases reported, customers gave the software their own passwords. That is where the trouble starts.

The Customer Says Yes. But What Is the AI Doing?

Unnamed sources told Axios that banks had detected unauthorized AI agents inside customer accounts, logging in with the customer’s permission and credentials. Most banks require the account holder to log in personally, so an agent using a customer’s password can break the bank’s rules even when the customer approved it.

In some cases, the agents contacted bank call centers or appeared to act independently of what the customer asked. The report does not say any agent moved money.

Banks cannot always tell the difference. One executive said they struggle to know whether an agent is authorized or has “gone rogue.” A customer may approve a tool for one job, and the bank has no easy way to confirm it stays inside that permission. Many banks are moving toward read-only access, which lets an approved tool see balances and history without moving money.

Wall Street’s Message: Our Systems, Our Rules

JPMorgan Chase CEO Jamie Dimon addressed it on the bank’s earnings call this month. “We’re not going to allow just any outside agent to come inside the company,” he said. “You don’t want rogue agents.”

Charles Schwab CEO Rick Wurster said the firm is blocking agents “because we want to have control over authentication and the experience.” He has also described “tens and twenties” of attempts by major AI agents to access client funds.

Morgan Stanley has blocked third-party agents, according to a person familiar with the decision, though a spokesperson said the bank has not seen “rogue” agents on its platforms. Citi says it does not permit third-party AI agents to log into customer accounts. Customers may love these assistants. The banks are not ready to give them a human’s access.

From a Security Concern to a Scary Headline

Florida Gov. Ron DeSantis reacted to the report on X. “Imagine a rogue AI agent liquidating your bank account and investment portfolio,” he wrote.

It is an alarming picture, and it is a hypothetical one. The report does not establish that any rogue agent has drained an account. Some headlines used similarly stark wording, which can sound like something that has already happened. Detecting agents that break a bank’s access rules is a long way from agents taking control of money.

The Fine Print Could Matter More Than the AI

The biggest question may be who pays when something goes wrong. Bank of America CEO Brian Moynihan put it bluntly: “When you give the agent your credentials, it is you authorizing the payment.”

Wells Fargo’s online access agreement, effective Nov. 24, treats an agent’s transactions as authorized unless the customer notifies the bank, and places the risk of granting agent access on the customer. Contract language does not settle every legal dispute, and how an unauthorized transaction is treated can depend on the circumstances and applicable law. Reading your own bank’s agreement is the practical first step.

What You Can Do

•      Do not give your banking password to an AI assistant. A Chase spokesperson said customers “should not give their usernames and passwords to anyone.”

•      Read your bank’s access agreement. Find out how it treats third-party agents and transactions made with shared credentials.

•      Use your bank’s official app or website. Some banks, including Citi, allow limited read-only access through approved open-banking providers.

•      Review your transactions regularly. If something looks wrong, call the number on your card or on the bank’s official website. Do not use contact details from an unfamiliar app or message.

•      Do not assume an AI tool’s promises override your bank’s rules. An assistant that says it can pay your bills has not shown that your bank allows it.

The Bottom Line

AI agents do not have to hack a bank to cause trouble. When a customer gives software a password, the bank loses the ability to know who is acting. The reporting describes banks restricting outside AI access. It does not show AI emptying accounts or locking people out.

The open question is who is responsible when an AI does something its owner never intended. Until that is settled, treat your banking password like the key to your house.

Sources:

•      Axios, “Scoop: Unvetted AI agents are accessing U.S. bank accounts” (Oct. 22, 2026)

•      Reuters, “US banks are blocking third-party AI agents access to customer accounts, sources say” (Oct. 22, 2026), via Investing.com

•      Fortune, Jamie Dimon on AI agents and JPMorgan (Oct. 15, 2026)

•      Ron DeSantis, post on X (Oct. 22, 2026)

•      Decrypt, “Major Banks Find Rogue AI Agents Inside Customer Accounts” (Oct. 2026)

USA Today, “‘Window of vulnerability’: Banks move to restrict AI agent access” (Oct. 23, 2026), via Yahoo Finance