Major financial institutions are issuing warnings to customers regarding the use of AI agents for account management. Banks emphasize that actions taken by AI tools, including unauthorized or erroneous financial transactions, may be legally considered as authorized by the user, leaving the customer liable for any resulting losses.

Ally Bank, JPMorgan Chase, and Citi have sent notifications or updated their platforms to clarify that AI agents can hallucinate or misinterpret commands. JPMorgan Chase specifically noted that transactions performed by AI, if misused, might be irreversible. Citi has added banners to its mobile app and website to alert users about potential risks.

The rise of autonomous AI agents, such as Meta's Muse, OpenAI's Dots, and Instinct AI, has prompted retailers to adjust their terms. In March, Target updated its user agreement to classify AI-driven purchases as authorized transactions. In September, Amazon banned Meta's Muse from its platform, citing security and reliability concerns.

Banks are urging customers to avoid sharing sensitive credentials like passwords or security codes with third-party AI tools. Wells Fargo representatives stated that no technology should act on behalf of a client without strict security measures. JPMorgan Chase CEO Jamie Dimon confirmed that the bank restricts external agents from accessing its internal systems.

The trend of issuing warnings has extended to technology companies. Apple has introduced risk alerts in the latest macOS beta versions, which notify users when they grant system-level access to AI agents or applications, advising caution before proceeding with such authorizations.