Banks warn customers: If your AI agent gets hacked and steals your money, you may be on your own

Six major banks are changing their terms for customers who let outside AI tools access their accounts. Citi, PNC, Citizens, Wells Fargo, JPMorgan Chase and Bank of America have all moved to limit their responsibility when third-party AI agents handle banking tasks.

Citi says if an AI agent exposes your login credentials and a thief uses them to move money, the bank may treat those transactions as authorized. Wells Fargo says customers use outside AI agents at their own risk. Its new terms take effect November 24. Customers are also told to keep written proof if they revoke an agent’s access.

So you give an AI tool access to your bank account, it gets hacked, and the bank may tell you the resulting theft is your problem.

There is a catch for the banks: federal Regulation E protects consumers from certain unauthorized electronic transfers. Whether these new contract terms can get around those protections has not been settled.

https://techcrunch.com/2026/11/18/more-banks-says-customers-using-ai-agents-are-on-their-own-even-if-they-get-hacked/

Grok factcheck:
Major banks including Citi, Chase, Bank of America, PNC and Citizens have updated terms shifting liability for third-party AI agent actions (even post-hack or credential leak) onto customers.

Citi Online User Agreement (Section H): https://citi.com/tandc/view-printable-version

Chase Digital Services Agreement: https://static.chasecdn.com/content/dam/legal-agreements/library/en/dsa1_la/versions/dsa1_la.pdf

PNC Digital Services Agreement: https://pnc.com/content/dam/digital-assets/en/documents/digital_services_agreement_UPDATE.pdf

Wells Fargo Online Access Agreement update: https://wellsfargo.com/mobile-online-banking/updates/

BofA and Citizens notified customers by email.

Got a news tip or correction? Let us know

If you got something out of this, please chip in to keep this site running, or subscribe to go ad-free.

10 views