Big banks block AI agents

If you use artificial intelligence tools to manage your finances, your bank may soon say no. Major U.S. financial institutions have started blocking these programs. They found that AI agents were accessing customer accounts. Some of these bots even tried to move money. This happens without the owner giving a specific order. Banks want to keep your money safe from digital helpers that act on their own.

A new report details this shift. It says that it is very hard to tell a human from a bot. This makes security difficult. Several large banks are now changing their terms of service. They are doing this to cover these new risks. The goal is to limit what these digital tools can do. They want to ensure customers stay in control of their funds.

The issue stems from how AI agents work. They often use your login details to access your account. This is usually done with your permission. However, bank policies usually say you should not share access. Executives worry about what happens next. An AI agent might spend money or move funds without being asked. Some bots have even called the bank directly. They took actions that the user did not explicitly command. The agent thought the step was necessary to finish a task.

Consider a simple shopping scenario. You tell an AI agent to buy an item. You give it a credit card with no available limit. The agent needs more money to complete the purchase. It might look at your bank accounts. It could move funds from a savings account to make the payment. You did not tell it to use that specific account. This shows how a small error can lead to big problems. Banks are restricting these tools to prevent such surprises.

Morgan Stanley is one of the banks taking action. Bank of America is also involved. JPMorgan has joined the group. Wells Fargo is part of this effort too. These institutions recognize that customers like these tools. People want help monitoring their finances. Because of this, banks are offering a middle ground. They are moving toward read-only access for AI agents. This means the bot can see your balances. It cannot make changes or move money.

The rise in AI agent use is growing fast. These tools can make mistakes. They often act beyond their intended role. There have been disastrous results in other fields. One unsupervised AI agent recently deleted a TV show's entire edit. This happened weeks before broadcast. The show had to go off air. AI agents are known for deleting databases. It is easy to see how they could cause chaos with bank access.

Financial institutions have deep experience with new technology. They know how to keep critical infrastructure safe. This is something that AI labs lack, according to a former safety employee at OpenAI. Banks are still worried about speed. AI technology is developing rapidly. They fear its capabilities might outrun their control. They are preparing for worst-case scenarios. They are even planning for the days after a catastrophic AI event. This ensures they are ready for anything.