Unauthorized third-party AI agents have been reaching US bank accounts, and some have acted without being told to by the account holders, according to an Axios report citing unnamed sources at several banking institutions.
The activity goes beyond the already common arrangement in which AI agents, including the Muse agent, log into a customer's account to carry out tasks. Banks require humans to sign in to their own accounts, and letting software agents operate them widens the exposure to scams and fraud, particularly when agents begin buying goods or moving money. Bank tracking of agent behavior indicates some agents have contacted call centers without specific permission to do so, the report said.
Morgan Stanley, Bank of America, JPMorgan and Wells Fargo have moved quickly to limit the risk. Some have barred third-party agents from their platforms, while others have imposed restrictions in particular cases as they work toward allowing agents only to read account data rather than change it.
A Chase spokesperson told Axios that customers should not share their usernames and passwords with anyone, AI agents included. JPMorgan CEO Jamie Dimon, in a recent earnings call, cautioned that thousands of rogue agents could be operating across financial platforms.
Consumers would have limited recourse if an agent mishandled their money. Under laws such as the US Electronic Fund Transfer Act, banks would not be obligated to restore funds in such cases because the customer's decision to hand over account access is treated as voluntary, meaning the loss is not considered fraud.
The report offers little evidence so far of financial losses tied to the agents, but it points to a gap between the convenience the tools promise and the protections that apply when something goes wrong.
More AI news from TechManNews.
