What the report found
Major U.S. banks are detecting activity from third-party AI agents inside their customers' online accounts, according to a new report from Axios. Some of the country's largest financial institutions say customers have given personal AI agents access to their financial accounts, even though the banks have not vetted those tools.
Madison Mills, Axios's senior AI reporter, told CBS News that in many cases customers are simply giving a username and password to an AI agent. She said that goes against a bank's terms of service. In other cases, she said, Axios found evidence that appeared to show agents taking actions in accounts that were not authorized.
One example of behavior a user never approved was an AI agent calling a bank with a complaint without first consulting the account holder. Mills described the situation as "a nightmare of AI agent activity for banks."
Not a case of rogue agents
Mills stressed that the agents were not going rogue or hacking accounts and draining funds. According to her, they were largely following user instructions, but without the bank's permission to transact on a human account holder's behalf.
She also said the issue is industrywide, affecting dozens of financial institutions as they race to curb AI agents' access to customer accounts.
White House adviser weighs in
White House AI czar David Sacks responded to the report in a post on X, drawing a line between personal AI agents that help manage bank accounts and hackers using AI for malicious purposes. He said there is "a world of difference" between user-authorized agents, naming Muse, Instinct and Grokbot, and "true unauthorized/rogue agents that could constitute cyberattacks."
Sacks said most of the activity in the Axios story appears to involve customers sharing credentials with agents so the agents can act for them. He said this may violate bank terms requiring humans to log in themselves, but "these are in no way cyberattacks."
He argued banks should support user-authorized agents. If banks want to limit access, he said, they could offer "read-only" permission. An agent with that access could view spending and other activity, such as judging whether a user spends too much on dining out, but could not make purchases or take other actions. Sacks said full write access with user credentials is "a different risk category and should be approached carefully."
How banks are responding
A Chase spokesperson told CBS News that customers should not give their username and password to anyone. The spokesperson said the bank supports secure, customer-authorized data sharing through open banking providers.
JPMorgan Chase CEO Jamie Dimon raised concerns about AI agents acting in customer accounts on the company's Oct. 13 earnings call. "We're not going to allow just any outside agents to come in inside the company because it's just too risky for clients and for JPMorgan," he said.
Charles Schwab's CEO said earlier this month that the company has been blocking unauthorized agents, according to Reuters. A Morgan Stanley spokesperson said the firm had not observed any unauthorized AI agent activity on its platforms. Bank of America and Wells Fargo did not immediately respond to CBS News' requests for comment.
