KEY POINTS
- Citi updated its Online User Agreement so customers are responsible for activity carried out by third-party AI agents they authorize.
- JPMorgan and Bank of America announced similar policies last month, and PNC changed its terms on the day of the report.
- Citi's terms also make customers liable if an AI tool leaks their credentials or is hacked.
Citibank, described by TechCrunch as the third largest commercial bank in the United States, has changed its terms of service to say that customers who use third-party AI tools for personal finance carry the responsibility if something goes wrong. That includes cases in which the AI tool is hacked.
AI agents, from models built by leading labs to purpose-built money-management apps, can be given the login details for bank accounts and other online financial portals. The aim is to let them analyze spending and investments, give advice, and even make payments. Banks, however, are wary of automated actors working inside their systems, and they say they will not reverse transactions made by such assistants, whether the action was intended or not.
A Citi spokesperson told TechCrunch that the bank wants customers to understand the risks of sharing credentials or authorizing outside AI tools to reach their accounts. The company said it updated its Online User Agreement to make clear that customers who grant access to third-party AI agents are responsible for activity conducted through them.
According to TechCrunch, Citi is not the first to take this position. JPMorgan and Bank of America, the two largest banks in the country, announced comparable policies last month, and PNC, the eighth largest, revised its terms on the day the article was published. Citizens also emailed customers this week with a similar warning, according to an email seen by the outlet.
Citi's update goes a step further than the others described in the report. It states that if an AI tool leaks a customer's credentials or is compromised, the customer is responsible for whatever hackers then do with the account.
The report notes that AI startups see personal finance as a way to add value and earn a share of the fees the sector generates. But customers who are already cautious about giving an AI agent access to email or chat history face a bigger trust question when the account is a bank or brokerage. TechCrunch pointed to Anthropic's personal finance tool, launched earlier this month, which is described as read-only and unable to move, manage or invest money.
Security is not the only concern for banks. TechCrunch said they also see AI agents as a competitive threat, since the tools could help consumers avoid unnecessary fees or low-interest deposit accounts that support bank margins. Markets reacted briefly to that prospect. The report adds that Citi's terms say customers can use AI agents provided by the bank itself without worrying about extra liability.
TechCrunch described the banks' stance as part of a wider pattern, in which AI labs promise their agents will change how people deal with service providers and those providers push back. It cited examples from car-hailing to food service, and said how these disputes are settled will help determine which AI investments pay off.