Bank of America Warns You May Be Liable for What Your AI Agent Does

The bank says customers are responsible for transactions made by AI agents they give access to, raising fresh concerns about agentic shopping.
We receive compensation from the products and services mentioned in this story, but the opinions are the author's own. Compensation may impact where offers appear. We have not included all available products or offers. Learn more about how we make money and our editorial policies.

AI agents have grown by leaps and bounds. They can now find the best deals for you, book flights and hotels, fill out forms, and even execute transactions on your behalf. But to do that, many of these agents ask for your login credentials, and that’s where banks are drawing the line.

Bank of America has warned its customers against allowing AI agents access to their banking accounts. In case of any financial mishap, the liability could ultimately fall on you, leaving you responsible for the losses.[1]

Here’s everything you should know about the latest Bank of America warning and the steps you should take to stay safe while using AI agents.

In this article
What did Bank of America say
The risks of giving AI agents access to your bank account
How to stay safe when using AI agents
The bottom line

What did Bank of America say

Bank of America warns customers against sharing their usernames and passwords with third parties, including AI agents that offer to act on their behalf. The warning was sent in an email titled “Protecting your finances in the AI World” and shared by Brian Toporek on Bluesky.

This comes just days after the bank published a joint report titled “Building Trust in Agentic Commerce” alongside ASB Bank, Capital One, Commonwealth Bank of Australia, ING, and NatWest, outlining steps to make AI-powered shopping safer.

Then there’s the question of liability. If something goes wrong, who should take the fall? Should it be the customer who shared their login credentials, the makers of the AI agent, or the banking institutions? Bank of America, in its email, has explicitly stated that customers will be responsible for activity resulting from such AI access.

The risks of giving AI agents access to your bank account

While there are several AI tools that can analyze your spending transactions and help you draft a budget, some even offer to authorize and execute transactions on your behalf. However, there’s always a possibility of an AI assistant executing transactions without your knowledge or approval.

We’ve already seen this happen. Geoffrey Fowler, a Washington Post columnist, asked an AI agent to find cheap eggs. The agent, however, went on to spend $31.43 on a dozen eggs using Fowler’s credit card, all without any instructions or approval.

Plus, banks are also struggling to identify whether agents are legitimately authorized by the account owner or are being controlled by malicious threat actors. This is why several major U.S. banks have started restricting the use of such AI agents.

Morgan Stanley has started blocking such agents, Citi is actively restricting access, and JPMorgan Chase has discouraged customers from sharing their logins with such AI agents. Your bank may think that an unauthorized party is raiding your account and lock you out of your own account.

Lastly, there’s also the risk of an AI agent failing to identify a fake shopping app and handing over your personally identifiable information, such as credit card details, email address, or home address, to threat actors who can then use it to orchestrate financial fraud schemes, send very convincing phishing messages, or even attempt identity theft.

Guardio Labs ran some controlled tests on Perplexity’s Comet browser, which included a fake Walmart store and a live Wells Fargo phishing page. The agents overlooked the obvious telltale signs that the store was fake, completed the purchase, and handed over payment details. This is exactly the kind of AI agent behavior that Bank of America has warned customers against.

That said, users still have protections under federal law, including the Electronic Fund Transfer Act and Regulation E, in cases of genuinely unauthorized electronic transfers. However, the extent of these protections can depend on the circumstances, particularly when customers have voluntarily shared their login credentials with a third party, such as an AI agent.

How to stay safe when using AI agents

You should limit the amount of information you share with AI agents. While it's fine to use them to analyze your finances, spending patterns, and habits, or even draw up your weekly and monthly budgets, make sure you're the one who authorizes the final transactions.

Additionally, consider using a reliable third-party antivirus program with real-time web protection to flag suspicious pages and links before you or your AI agent interact with them.

The bottom line

With Bank of America’s latest warning against giving away your bank passwords to AI agents, it’s high time to think about the extent of control you’re willing to hand over to AI assistants.

In case of any financial or transactional mistakes, the liability could ultimately fall on you, as Bank of America has clearly stated. Other banks, such as Morgan Stanley, Citi, and JPMorgan Chase, have also started restricting or discouraging the use of AI agents on their banking portals.

You may still continue to use these agents to analyze your finances, spending patterns, and draft budgets. However, it’s recommended that you refrain from allowing these agents to execute transactions on your behalf.

4.8
Editorial Rating
See Price
On Aura Identity Theft's website
2026 Editors’ Choice
Best Overall Identity Theft Protection Service
Identity Protection
Aura Identity Theft
  • ID theft protection that monitors your SSN, bank accounts, credit cards, and brokerage and retirement accounts for suspicious activity
  • Every plan includes the full feature set, so no additional cost to unlock monitoring, insurance, or restoration
  • Bundles data removal with identity theft protection, antivirus, VPN, and a password manager in one subscription

Author Details
Krishi Chowdhary specializes in digital privacy, cybersecurity, and consumer technology. He has written extensively on online privacy tools and broader cybersecurity topics, including online scams, data breaches, age verification, and emerging digital threats. Krishi believes technology reporting should empower readers, not confuse them, and is committed to making even the most technical subjects easy to understand without compromising on accuracy or depth. His work has appeared in leading technology publications, including CNET, ExpressVPN, and TechRadar, where he has covered topics ranging from cybersecurity incidents and privacy product announcements to artificial intelligence and major technology news

Citations

[1] Bank of America’s email shared by Bryan Toporek on Bluesky