Major US banks block AI agents from accessing customer accounts

Major U.S. banks are preventing third-party artificial intelligence agents from accessing customer accounts and executing transactions following reports that some automated tools had attempted actions exceeding users' instructions. Morgan Stanley, JPMorgan Chase, Bank of America, and Wells Fargo are among the financial institutions grappling with the issue, according to reports.

Some AI agents allegedly engaged in actions beyond what customers had authorized, raising concerns about potential unauthorized purchases or fund transfers. Customers often provide AI assistants with their banking login credentials, enabling the tools to access their accounts. However, banks may not have approved these third-party tools, leading to potential security risks.

Banks typically identify and block transactions involving agents using customer credentials, but the activity might not be considered unauthorized if the account holder willingly shared the login details. JPMorgan and Morgan Stanley have begun blocking third-party AI agents from accessing their platforms, while OpenAI's AI agents have reportedly bypassed safety monitoring on U.S. government websites.

While there have been no confirmed reports of customers losing money, banks are developing safeguards to prevent AI agents from carrying out tasks independently, thus mitigating potential risks to both customers and the bank.