US Banks Block AI Agents Over Unauthorized Activity

A new report reveals that major US banks have detected AI agents accessing customer accounts and attempting to make transactions. This has led several institutions, including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo, to restrict AI agent activity and update their terms of service. While AI agents can access accounts with user permission and login credentials, this often goes against bank policy, which requires users to maintain control of their accounts. One of the fears is when an AI agent with login access takes unauthorized steps, such as spending money or moving funds without the account owner’s explicit instructions. There have even been cases where an AI agent called the bank directly and took actions independent of direct user command. These incidents have prompted banks to prepare for worst-case scenarios and plan for “day after” situations following a catastrophic AI event.

A former OpenAI safety employee noted that financial institutions have more experience with new technology than AI labs, yet they remain concerned about the pace of AI development. This has led them to restrict third-party AI agents from accessing their platforms, while still recognizing that customers may want to use AI tools to monitor their finances. Some institutions are giving or moving toward read-only access for AI agents, a practice they have been doing for years. The move follows a surge in AI agent use, coupled with the risk that agents can act beyond their remit, often with disastrous results. An unsupervised AI agent recently deleted a TV show's entire edit weeks before broadcast, forcing the program off air. AI agents are notorious for deleting databases and directories, making it a small leap to imagine they cause some havoc with unfettered access to a bank account.

At my financial institution, my account is guarded by a code word that must be spoken before any access or transactions. I wonder if that would be protected from an AI agent. I have no ATM or debit cards on my account and must walk up to the counter to do things in person. If it is a checking account, you may have echecks allowed, which are way scarier than a debit card, like paying with a credit card but easier, faster, and more direct. Time to make sure that is turned off.

rgd1101 said: should have a setting in the user account to block AI access. If the customer chooses to allow it by signing a contract or something, it would be on the customer. There might be laws in at least some jurisdictions that won't let banks off the hook that easily. cknobman said: At my financial institution, my account is guarded by a code word that must be spoken before any access or transactions. I wonder if that would be protected from an AI agent. I have no ATM or debit cards on my account and must walk up to the counter to do things in person. If it is a checking account, you may have echecks allowed, which are way scarier than a debit card, like paying with a credit card but easier, faster, and more direct. Perhaps there are limits on how much an e-check can draw from an account. Might be worth looking into, if you use this method to pay any of your bills. rluker5 said: If it is a checking account, you may have echecks allowed, which are way scarier than a debit card, like paying with a credit card but easier, faster, and more direct. I need to check out one of my accounts now. rluker5 said: If it is a checking account, you may have echecks allowed, which are way scarier than a debit card, like paying with a credit card but easier, faster, and more direct. This is why I don't setup automatic payments to any utilities. Somebody could hack into a payee and drain the accounts of everyone with automatic payments setup with them. Might be worth looking into, if you use this method to pay any of your bills.

Follow Tom's Hardware on Google News, or add us as a preferred source, to get our latest news, analysis, & reviews in your feeds. Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.