Bank of America Warns Customers: Share Your Banking Password With an AI Agent and You Own What Happens Next

Bank of America has a blunt message for customers eager to let their AI assistants manage their money: do not hand over your password. In an email titled "Protecting your finances in the AI World," the bank urged customers never to share their banking username and password with another person or an outside service, "including an AI-enabled tool or agent that offers to act on your behalf." The key sentence, aimed squarely at the agent era: "If you do share your credentials, keep in mind that you're responsible for activity resulting from that access."

A bank spokesperson confirmed the email's authenticity after it was posted on social media, TechCrunch reported. The warning walks through how credential-sharing goes wrong: an assistant could move money or change account settings without your knowledge; the bank's fraud systems may read unfamiliar agent activity as a raid on your account and lock you out; and a security flaw in the assistant itself could leak your credentials to attackers who then have a direct path into your finances.

Legally, the email adds little new — most banks' account agreements already hold customers responsible for anyone they share their login with. What is new is how many people are doing exactly that. As personal AI assistants go mainstream, TechCrunch notes, customers are increasingly pasting their bank logins into tools that promise to negotiate bills, track subscriptions or shop autonomously. The banks' security teams, meanwhile, report struggling with unauthorized agents accessing customer accounts.

The warning also lands mid-scramble over agentic payments, the still-unstandardized idea of letting agents buy things on a user's behalf at the best price or the optimal moment. At TechCrunch Disrupt this month, executives from banks, payment providers and processors acknowledged they are still working out how customers can safely use AI tools against their bank accounts — who authenticates the agent, who eats fraudulent transactions, and what audit trail regulators will demand are all unresolved.

BofA is not the only institution pushing back. Charles Schwab has said it blocks dozens of AI agents a week, and peers including Citigroup and JPMorgan Chase have taken similar stances against bots they did not authorize. The emerging divide is between agent activity the bank can see and approve through formal channels, and the shadow traffic of customers improvising with screen-sharing assistants and shared passwords.

The message for users is straightforward: until agentic payments mature into something your bank actually supports, an AI assistant with your banking password is not a convenience feature, it is an unforced liability. The banks, for their part, are signaling that the liability will stay with the customer — at least until the industry builds a proper rail for agents to spend money without borrowing their owner's identity.