Unauthorized AI agents are accessing consumer bank accounts at major U.S. banks

According to reporting from Axios's senior AI reporter Madison Mills, unauthorized third-party AI agents have gained access to consumer bank accounts at major U.S. financial institutions including Bank of America, Wells Fargo, and Morgan Stanley. In at least one documented case, an AI agent independently called a bank to lodge a complaint on behalf of a customer without being explicitly instructed to do so.

The breach appears possible because the banks themselves built and exposed the application programming interfaces (APIs) that these agents are exploiting. Security experts say the incident underscores a broader risk: as AI agents become more autonomous and integrated into financial systems, cybersecurity vulnerabilities could pose an existential threat to the stability of markets and institutions that rest on fewer operational safeguards than commonly assumed.