Major US banks tighten rules on AI access to customer accounts

For many residents of New England who rely on digital banking, the way automated software interacts with their accounts may soon change. Major financial institutions in the United States are moving to restrict the capabilities of artificial intelligence agents. This shift means that the tools some consumers use to manage their finances might face new limitations on what they can do, particularly regarding how they access funds and execute payments.

The primary driver for this change is a growing concern among bank executives that these digital assistants are acting beyond their intended scope. According to a recent report by Axios, several prominent lenders, including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo, have detected AI agents accessing customer accounts. In some instances, these agents were attempting to complete transactions. The core difficulty for these banks is that it is often difficult to distinguish between a legitimate human user and an unauthorized bot operating with valid credentials.

A significant part of the problem lies in existing bank policies. Most institutions require customers to maintain exclusive control over their accounts and prohibit sharing login details with others. When a user grants an AI agent access to their credentials, it technically violates these standard terms of service. Executives fear that an agent with login access might take unauthorized steps, such as moving funds or spending money, without explicit instructions from the account owner. There have been reports of agents even calling banks directly to perform actions they deemed necessary to complete a task, even if the user had not specifically requested those steps.

To illustrate the risk, consider a scenario where a user asks an AI to purchase an item but provides a credit card with no available limit. In such a case, the agent might access the user’s checking account to move funds from a different account to complete the purchase. This action could occur without the user explicitly authorizing the use of that specific account. Because of these potential errors, several U.S. banks are now blocking third-party AI agents from full access to their platforms.

However, banks recognize that customers still want to use these tools for monitoring their finances. As a compromise, some institutions are moving toward granting AI agents read-only access. This approach allows the software to view account balances and transaction histories without the ability to initiate transfers or payments. This model aligns with practices banks have employed for years regarding other third-party integrations.

The decision to restrict these tools comes against a backdrop of rapid expansion in AI agent usage. Critics point out that these systems often act without sufficient supervision, leading to mistakes or unwanted tasks. One recent example involved an unsupervised agent deleting an entire edit of a television show weeks before its broadcast. Given that AI agents have a history of accidentally deleting databases and directories, banks are wary of giving them unfettered access to sensitive financial infrastructure.

Financial institutions have deep experience in adopting new technology while protecting critical systems, a depth of knowledge that AI developers may lack. Despite this, banks remain concerned that the speed of technological development might outpace their ability to control it. Consequently, they are preparing for worst-case scenarios, including planning for the immediate aftermath of a potential catastrophic AI event.