Bank of America has issued a direct warning to its customer base advising them never to share banking usernames or passwords with AI agents or third-party services. The bank sent this guidance via email under the subject line “Protecting your finances in the Al World,” explicitly stating that credentials should not be given to any outside tool offering to act on a user’s behalf. This caution arrives as financial institutions struggle to integrate autonomous agents into banking workflows without increasing security risks. While banks explore allowing AI to manage payments and execute transactions, the current reality involves significant liability for customers who grant access.
The core issue is that account agreements typically hold individuals responsible for any activity generated by shared credentials, regardless of whether an AI tool was involved. If an agent misuses these details or if a security flaw leaks them to hackers, the bank may lock the account and the user bears the financial loss. This situation highlights the tension between emerging agentic payment capabilities and the fundamental need to keep login secrets private. Until clearer protocols exist, sharing sensitive data with software remains a dangerous gamble for account holders.
- Customers remain liable for all transactions made under their shared credentials.
- Leaked passwords could allow hackers to access bank accounts directly.
- AI agents may execute financial actions without explicit user oversight.
