Banks Block AI Agents From Unauthorized Account Access

AI agents have moved from chatbots to doing actual digital work: booking flights, filling out forms, logging into accounts. That shift just ran into a wall at some of the largest banks in the country. JPMorgan, Bank of America, and Morgan Stanley have all caught AI agents operating inside customer accounts using the customer's own login credentials, according to a report from Axios.

The issue isn't just that agents are showing up uninvited. It's that banks can't always tell which ones were authorized by the account holder and which ones are acting on their own, taking actions nobody asked for. That distinction matters a lot when the agent in question can move money, call customer support, or make purchases without a human checking first.

Banks have spent years building fraud detection systems designed to spot human behavior patterns, login times, device fingerprints, typing habits. AI agents break those assumptions entirely, since they can mimic a legitimate user's credentials while behaving nothing like them. What's happening inside these institutions now is a scramble to figure out where the line between helpful automation and financial risk actually sits.