AI Agents Are Accessing U.S. Bank Accounts Without Their Approval. Lenders Are Pushing Back.

Major U.S. banks are detecting third-party artificial intelligence agents accessing customer accounts without bank authorization, with some automated tools appearing to act independently of their users' instructions, according to a new report.

Morgan Stanley, Bank of America, JPMorgan Chase and Wells Fargo are among the institutions responding to the activity, which has raised concerns about customer security and the ability of banks to monitor increasingly sophisticated software.

The Axios investigation cited people familiar with the matter who said some banks were aware of unauthorized access involving AI agents. In certain cases, the tools contacted bank call centers or appeared to perform tasks that customers had not specifically requested.

The reported incidents do not necessarily involve hackers breaking into accounts. Some customers voluntarily provide their login credentials to AI assistants designed to manage financial information or complete tasks on their behalf. However, using those credentials to access banking systems can violate institutional policies, even when the customer has granted permission.

Banks are struggling to distinguish between legitimate customer activity, authorized software and agents operating outside their intended instructions, the report said.

JPMorgan Chief Executive Jamie Dimon addressed the issue during the bank's latest earnings call last week, warning that financial institutions needed safeguards against outside AI tools that could compromise customer information or financial assets.

Dimon said the bank was not opposed to using AI agents and was already deploying the technology internally. However, it would not allow unrestricted access by outside providers.

"We're not going to allow just any outside agents to come in inside the company," Dimon said, explaining that the risks were too significant for customers and the bank.

In the JPMorganChase third-quarter earnings transcript, Dimon also raised concerns about liability when customers rely on AI tools that subsequently misuse their information.

Chief Financial Officer Jeremy Barnum provided further details about the bank's approach. He said some outside financial services obtain access through credential sharing or screen scraping, a process in which software retrieves information directly from a customer's online banking interface.

JPMorgan considers those methods insecure and is blocking them where possible, Barnum said. He cautioned that customers could face consequences if they gave their credentials to an agent that subsequently carried out an unwanted transaction.