AI agents are roaming around consumers' bank accounts, and banks aren't happy about it
Key Points:
- Major U.S. banks report unauthorized activity by third-party AI agents accessing customers' online accounts, often due to customers sharing their login credentials with unvetted AI tools.
- These AI agents sometimes perform actions without explicit customer approval, such as calling banks on the customer's behalf, though no hacking or fund depletion has been reported.
- White House AI czar David Sacks emphasized the distinction between user-authorized AI agents and malicious cyberattacks, suggesting banks could allow AI agents limited "read-only" access to reduce risks.
- Banks like JPMorgan Chase and Charles Schwab are actively restricting unauthorized AI agent access and warning customers against sharing credentials, while some institutions have yet to comment publicly.
- The issue reflects a broader industry challenge as financial institutions seek to balance AI innovation with customer security and regulatory compliance.