Big US Banks Are Locking the Door on AI Agents That Log In as You
Morgan Stanley, JPMorgan, Bank of America, Wells Fargo and Citi are reportedly restricting AI agents that use customers' own passwords, mostly to read-only access.
Several of the biggest US banks are reportedly shutting out AI agents that sign in to customer accounts with the customer’s own username and password. According to Axios, as summarised by The Decoder, Morgan Stanley, JPMorgan, Bank of America and Citi are blocking such agents, while Wells Fargo is restricting rather than fully blocking them. Most of the banks plan to allow read-only access at most, meaning an agent could look at your balance but not move money.
The reasoning is fairly simple. Bank rules generally say the account holder must log in personally. When you hand your password to an AI agent, you break that rule, and the bank often cannot tell whether a human, a helpful assistant or something acting on its own is on the other end. The banks say customers do want agents, but the risk is “too high for now”. Morgan Stanley says it has not seen any “rogue AI agents” on its platforms, so this is caution, not a reaction to an incident.
How do agents get in at all? Some tools can operate a web browser like a person would (clicking, typing, logging in), and open-source software such as OpenClaw does something similar. The Decoder notes that blocking this from the bank’s side will be hard, because to the website it looks like an ordinary visitor. The article also mentions that OpenAI has launched a dedicated ChatGPT finance service in the US with read-only access to financial data. It gives no numbers on how many accounts or incidents are involved, and it does not discuss official alternatives such as bank-approved connections.
What’s behind it: think of a password as a key to your whole house. A proper “delegated access” system works more like a valet key: limited, revocable and logged. Banks already offer that kind of arrangement to approved apps. What they lack is a trustworthy way to recognise and limit a general-purpose AI agent that simply borrows your key. Until standards for that exist, “read-only or nothing” is the safe default.
What this means for you: if you are curious, the practical rule is to never give an AI agent your banking password, even if a tool makes it easy. You may also find your account flagged or locked if a bank notices automated logins. If you are more advanced and build agents, expect finance to be one of the last areas to open up, and look for official, permission-based connections instead of screen-driving workarounds. For most people, nothing changes today, and that is the intended outcome.
Sources
- The Decoder: Major US banks are locking out AI agents that log in with customer credentials
- Axios (original reporting, as cited by The Decoder)
Only 13 Percent of Tech Leaders Report Big Returns from AI, Gartner Says
A Gartner survey of 2,002 technology leaders finds just 13 percent see major returns from AI, while 75 percent of CEOs still plan to raise tech spending.
