Startups & Technology

Banks warn customers they bear risks of third-party AI agents

Banks warn customers they bear risks of third-party AI agents

AI agents can be given access to bank and other financial accounts to review spending and investments, offer advice, make transactions and pay bills. Banks say customers may have to bear the cost of activity carried out by these tools, whether it was intentional or accidental, and may not be able to get transactions reversed.

A Citi spokesperson told TechCrunch the bank updated its Online User Agreement to clarify the risks of sharing credentials or authorizing third-party tools. Customers who grant account access to AI agents, the spokesperson said, are responsible for what those agents do. Citi’s updated terms also place responsibility on customers if an AI tool leaks their credentials or is hacked and attackers use the account. JPMorgan and Bank of America announced similar policies last month; PNC changed its terms on the day of the report. Citizens also emailed customers this week warning that they were liable for actions by AI assistants, according to an email seen by TechCrunch.

The policies come as AI companies explore personal finance as a source of revenue, while consumers weigh the risks of giving software access to sensitive financial information. Anthropic’s personal finance tool, launched earlier this month, is billed as “read-only” and cannot move, manage or invest users’ money.

Banks also see AI tools as potential competitors. If agents help customers avoid fees or shift funds away from low-interest deposit accounts, they could challenge revenue streams that support bank margins. Citi’s terms distinguish its own AI agents, which customers can use without additional liability, from third-party tools. The dispute reflects a wider tension between AI companies seeking to mediate services and providers seeking to retain control over how customers reach them.

Share

Comments (0)

Leave a comment

No comments yet. Be the first!