Leading US banks have made a ‘horrifying’ discovery, uncovering evidence of unauthorized AI agents accessing customer accounts.
Companies and individuals alike have been deploying AI agents, giving users the ability to tap into agents to perform any number of tasks on their behalf. A popular task for AI agents is to handle financial matters. However, it appears that customers are unleashing their AI agents onto their bank accounts, despite those agents being unvetted and unauthorized.
According to Axios, some of the biggest banks in the US, including Bank of America, JPMorgan, Morgan Stanley, Wells Fargo, and others, are racing to contain the risks, and some have detected unauthorized AI agents accessing customer accounts. In many cases, it appears the agents are using valid login information that the customers provided. Unfortunately, while users may think it’s OK to use AI agents in this manner, giving an AI agent login information is against most banking guidelines.
To make matters worse, bank execs say AI agents may be taking actions the customer doesn’t expect, or that may not be in the customer’s best interests.
Bank executives described the situation as horrifying, with one official saying the industry has grave concerns about agents gaining access to customer accounts and taking steps that could impact user finances.
A Chase spokesperson said customers should not give their usernames and passwords to anyone. A Morgan Stanley spokesperson said the bank hasn’t observed instances of “rogue AI agents accessing our platforms.”
Further complicating the issue is that banks are struggling to differentiate between genuine human activity and that of AI agents. As a result, in the short term, some banks are opting to block AI technology from accessing their system altogether, while others are changing their terms of service to address concerns.
Over the long term, Axios says that most banks are moving toward a read-only access model that will allow users to give their AI agents access to their accounts without the risk of a rogue agent taking unwanted action.
The issue underscores the many ways AI is revolutionizing industries and activities far outside the normal tech sector, forcing a rethinking of how long-standing processes work.