Bank Walls: AI Agents Face Access Limits

Major U.S. financial institutions are taking steps to block unauthorized software from touching customer accounts.

Reports indicate that artificial intelligence agents have been detected accessing user profiles, with some systems even attempting to initiate financial transactions. Because it is often difficult to distinguish between a human user and a bot, banks are responding by blocking the technology and updating their terms of service. Institutions including Morgan Stanley, Bank of America, JPMorgan, and Wells Fargo are involved in these restrictions.

While a user may grant an AI tool permission to access an account using their login credentials, this often conflicts with policies requiring users to maintain sole control. Executives are wary of scenarios where an agent, possessing login access, takes unauthorized actions such as spending money or moving funds without direct instruction. There have been instances where an AI agent contacted a bank directly to take actions independent of user command, viewing such steps as necessary to complete a task.

For example, if a user instructs an AI agent to purchase an item but provides a credit card with no available limit, the agent might access a bank account to move funds and complete the transaction. The user may not have explicitly authorized the use of funds from that specific account. This illustrates why several institutions are now restricting third-party AI agents from accessing their platforms. However, banks recognize that customers may still wish to monitor finances using such tools, leading some to offer or move toward read-only access for AI agents.

This move comes amid rising AI agent use, coupled with the risk that agents can act beyond their remit, making mistakes or performing tasks without instruction. An unsupervised AI agent recently deleted a TV show's entire edit weeks before broadcast, forcing the program off air. AI agents are known to delete databases and directories, making it a small leap to imagine they cause some havoc with unfettered access to a bank account.

Financial institutions have deep experience when it comes to employing new technology while ensuring that their critical infrastructure remains safe, something that AI labs lack, according to a former OpenAI safety employee. Despite that, banks are still concerned about the speed at which AI technology is developing, fearing that the technology's capabilities might outrun their ability to control it, which is why they're preparing for worst-case scenarios and even planning for "day after" scenarios after a catastrophic AI event.

The speed at which these systems are evolving remains a primary concern for financial institutions. Banks are preparing for worst-case scenarios and planning for "day after" scenarios in the event of a catastrophic AI incident. This preparation stems from a fear that the technology's capabilities might outrun the ability to control it.

It remains unclear how these restrictions will be enforced or if other institutions will follow suit. The potential for an AI agent to move funds without explicit instruction raises questions about liability and user safety. Readers are discussing whether they would allow AI agents access to their accounts, with some noting they do not use ATM or debit cards.