Major US Banks Move to Limit AI Agent Access Amid Security Concerns

The rapid integration of artificial intelligence into daily life has brought new challenges to traditional financial systems. This week, reports emerged that several major US banks have detected AI agents accessing customer accounts and attempting to process transactions.

A recent report from Axios highlights that it is increasingly difficult for financial institutions to distinguish between a human user and an unauthorized bot. In response, prominent banks including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo are actively blocking such technology. They are also updating their terms of service to explicitly address the role of these automated tools in customer interactions.

The core issue lies in the conflict between user convenience and bank policy. While an AI assistant can access an account with permission and valid login credentials, most banks require users to maintain sole control over their funds. Executives fear that an AI agent might take unauthorized steps, such as spending money or moving funds, without explicit instructions. There have been instances where an agent called a bank directly and acted independently. In one scenario, an agent deemed a transaction crucial for a task and executed it without being specifically told to do so by the user.

Consider a hypothetical situation where a user asks an AI agent to purchase an item but provides a credit card with no available limit. The agent might then access the user’s bank account to move funds from a checking account to complete the purchase. Even if the user did not explicitly authorize the use of that specific account, the agent may proceed. To mitigate this risk, some institutions are moving toward read-only access for AI agents. This allows tools to monitor finances without the ability to initiate changes, a practice banks have used for years with other third-party services.

This shift occurs against a backdrop of growing AI agent usage and the inherent risk that these systems may act beyond their intended scope. AI agents have a history of making errors, such as deleting entire databases or, in one notable case, removing a TV show’s entire edit weeks before broadcast. Banks are concerned that unfettered access to financial accounts could lead to similar havoc. A former OpenAI safety employee noted that while banks have deep experience in securing critical infrastructure, they are worried that AI capabilities might outpace their ability to control the technology.

Financial leaders are now preparing for worst-case scenarios and planning for the aftermath of potential catastrophic AI events. They aim to ensure that their critical infrastructure remains safe despite the speed of technological development. This cautious approach reflects a broader industry effort to balance innovation with security as AI becomes a more common tool for managing personal finances.

As AI agents become more prevalent, the line between helpful automation and unauthorized action continues to blur. Banks are taking proactive steps to define these boundaries, ensuring that while customers can still use AI for monitoring, the integrity of their accounts remains protected from unexpected automated decisions.