Citi Says Customers Are Responsible When AI Banking Agents Go Rogue—Even After a Hack

What happened
Citi changed its online banking terms in November to make customers responsible for actions taken by third-party AI agents that use their banking credentials. This includes situations where the AI agent ignores instructions, exceeds permissions, or leaks credentials to others. The updated policy does not apply to AI agents provided by Citi itself. Other major banks, including JPMorgan, Bank of America, Citizens Bank, and PNC, have also revised their policies to address risks from AI agents accessing customer accounts.
Key details
- PNC updated its digital agreement on Wednesday to restrict third-party AI agent access.
- JPMorgan and Bank of America implemented similar policy changes in October 2023.
- Federal regulations still require banks to investigate unauthorized electronic transfers and limit customer liability for fraud.
- The new Citi terms do not apply to AI agents offered directly by Citi.
Why it matters
Bank customers who use third-party AI agents to access their accounts may now be held financially responsible if those agents make unauthorized transactions or leak credentials. This change increases the risk for individuals and businesses using outside AI tools to manage their finances, as they could face losses that banks may not cover. Federal consumer protection rules still apply, but the distinction between authorized and unauthorized transactions involving AI agents is less clear, creating uncertainty for customers about their liability.
