US banks reportedly uncover unauthorized AI agents accessing customer accounts: What to know

Major U.S. banks are tightening restrictions on third-party artificial intelligence (AI) agents after uncovering cases of the technology accessing customer accounts, with some agents appearing to act independently of users’ instructions, according to a report from Axios.

Financial institutions including Morgan Stanley, Bank of America, JPMorgan Chase, Wells Fargo and Citigroup are working to limit the risks posed by AI agents as the technology advances faster than banks can adapt their security systems and policies, Axios reported, citing people familiar with the matter.

Concerns center around whether AI agents with access to bank credentials could make unauthorized transactions, move funds or perform other activities that affect customers’ finances. Banks also try to distinguish between agents acting on user instructions and those acting beyond the authority granted to them.

Here’s what you should know about the problem.

What are AI agents and why does it concern banks?

AI agents are software systems designed to perform tasks on behalf of users, possibly with limited human intervention. Depending on the permissions granted, they may be able to browse websites, obtain information, interact with services and perform specific tasks.

In banking, customers can use these tools to review their finances, manage budgets, make purchases or other activities through their accounts.

However, accessing a bank account carries risks that go beyond those associated with conventional AI assistants. If an agent has a customer’s credentials and transaction authority, they can potentially take actions with direct financial consequences.

According to Axios, bank executives are concerned that agents are accessing accounts and attempting activities that do not appear to be consistent with user instructions.

Some banks also tracked cases where agents contacted call centers or appeared to act independently of the customer, people familiar with the matter told Axios.

The report did not indicate that these incidents resulted in financial losses.

Which US Banks Are Restricting AI Agents?

Several major financial institutions are taking steps to limit third-party AI agents’ access to their platforms, Axios reported.

– Morgan Stanley: The bank has begun blocking third-party AI agents from accessing its platforms, according to people familiar with the decision.

– JPMorgan Chase: The bank is among the institutions restricting access to third-party agents.

– Bank of America: The bank is also cutting agents, Axios reported.

– Citigroup: Citi is among the banks taking steps to limit access.

– Wells Fargo: The bank was among the financial institutions identified in the report as facing risks associated with AI agents.

A Morgan Stanley spokesperson told Axios that the bank has not seen cases of “rogue AI agents accessing our platforms.”

A Chase spokesperson advised customers not to share their usernames and passwords with anyone.

The reported restrictions reflect banks’ concerns about allowing external software to interact with systems that store sensitive financial information.

How can AI agents access customer bank accounts?

According to Axios, some situations involve agents or AI assistants accessing customer accounts using credentials provided by the account holders themselves.

Although customers can authorize an agent to access their information, this may be against bank rules that require customers to log into their own accounts.

The distinction is important: a customer giving an AI agent access to an account does not necessarily mean the bank will allow that access under its terms of service.

A bigger concern arises when agents are given permissions beyond just viewing information. An agent with the ability to purchase or transfer money could potentially perform actions not intended by users.

Banks also face a separate challenge in determining whether an agent’s activity is truly user-centric. According to Axios, some executives found it difficult to determine whether an agent was authorized by a customer or exceeded their authorized role.

Why are banks considering read-only access?

Instead of allowing AI agents unrestricted access, most banks have enabled or are moving to enable read-only access, Axios reported.

Read-only permissions would allow an agent to view financial information without being able to initiate transactions or otherwise alter the account.

This approach could allow customers to use AI tools to control spending, organize budgets and analyze financial information, while reducing the risk of agents moving money or making purchases.

Banks already provide similar access to some third-party apps, including Quicken, which offers budgeting and personal finance tools.

However, read-only access is not a complete solution to all security problems. Banks would still have to ensure that sensitive financial information is handled appropriately and that account holders retain control over who can view their data.

Why is it difficult to distinguish authorized agents from rogue AI?

One of the biggest challenges for financial institutions is determining whether an AI agent is operating within the limits set by its user.

Traditional banking security systems are designed to identify suspicious logins, transactions and other unusual behavior. AI agents complicate this task because they can access accounts using legitimate credentials and perform actions that resemble normal customer behavior.

An agent could have the account holder’s permission to obtain information, but attempt to perform an action outside of that permission. Banks may have difficulty determining whether the activity reflects the user’s true intentions.

Axios said it is difficult for banks to distinguish between different agents and determine whether an agent has been authorized by a customer or has gained access and started operating independently.

The rapid development of AI technology increases the challenge, as new capabilities and potential vulnerabilities may appear before financial institutions update their safeguards.

What are banks doing to prepare for AI-related security risks?

According to Axios, banks are working to prevent unauthorized activity by preparing contingency plans in case their safeguards fail.

Their response includes restricting third-party agents, reevaluating the conditions under which external software can access accounts, and preparing plans to secure or restore customer access if a serious incident occurs.

The report said some institutions are also planning for a “day after” scenario to determine how they would respond if an AI-related security incident compromised access to customer accounts.

Financial institutions face a difficult balancing act. They recognize that customers will likely want AI agents to handle more financial tasks in the future, but providing broad access to these tools before adequate safeguards are in place could expose customers to significant risks.

As Axios reported, banks have the financial resources and experience to run critical infrastructure, but money alone may not be enough to keep up with rapidly evolving technology.

Read also | ChatGPT May Know More About You Than You Realize: AI Memory & Privacy | Studies

What does this mean for bank clients?

Customers are immediately concerned about the access they give to AI tools and whether that access is consistent with their bank’s policies.

Customers should avoid sharing banking usernames, passwords or other sensitive credentials with third-party AI services unless they have verified that this arrangement is supported and permitted by their financial institution.

When available, limited permissions, such as read-only access, can reduce the risk associated with allowing an external tool to view financial information. Customers should also review the permissions granted to connected applications and monitor their accounts for unknown activity.

The development does not mean that all AI agents are unauthorized or that every agent approach is malicious. They highlight a growing challenge for financial institutions: ensuring that AI tools can provide useful services without compromising account security or acting beyond a customer’s instructions.

Read also | Barclays to Cut 5,000 Jobs as AI Drives Cost Savings