Citibank, the third-largest commercial bank in the U.S., changed its terms of service this week to tell customers who use third-party AI tools for personal finance that if anything goes wrong, it’s their problem, even if they get hacked.
From Frontier Lab models to dedicated personal finance tools, a variety of AI agents can be given credentials to bank accounts and other online financial portals, with the idea that they can analyze spending and investment decisions, provide advice, and even handle transactions and bill payments.
However, financial institutions are not satisfied with the presence of untrustworthy AI actors in their systems and are telling customers that they are not responsible for reversing intentional or unintentional transactions made by automated assistants.
“As third-party AI tools become more widely available, we want our customers to understand the risks associated with sharing their banking credentials or allowing those tools access to their accounts,” a Citi spokesperson told TechCrunch. “We recently updated our Online User Agreement to be more clear about these risks and reinforce that customers who choose to grant account access to third-party AI agents are responsible for the activities that occur through those agents.”
The country’s two largest banks, JPMorgan and Bank of America, announced similar policies last month. And PNC, the No. 8 bank, changed its terms today.
Citi’s update does more than just convey that the responsibility for everything the agency does lies with the customer. We also inform our customers that if the AI tool compromises their credentials or is hacked, they are responsible for anything the hacker does to their account. Banking giant Citizen similarly emailed customers this week, warning them that they were also at risk from the actions of its AI assistants, according to an email reviewed by TechCrunch.
As AI startups seek more revenue, entering personal finance presents an attractive opportunity for companies to add value to customers and tap into the fee streams generated in this space. But if consumers are already uncomfortable sharing email and chat records with AI agents, bank and brokerage accounts pose new trust challenges.
Anthropic’s approach is telling, as the personal finance tool it launched earlier this month explicitly claims to be “read-only” and “cannot move, manage, or invest your money.”
Beyond cybersecurity concerns, banks view entry into their franchises as a competitive threat. Markets briefly trembled in response to the prospect that AI tools would make it easier for consumers to avoid unnecessary fees and low-interest savings accounts that eat into banks’ margins. But if AI agents become a significant means of involvement in financial decision-making, banks will fight to control their channels. Citi’s terms specify that customers can use the AI agents it provides without worrying about additional liability.
The bank’s response is part of a broader trend among companies across all sectors. AI Labs says that when agents promise to revolutionize the way people interact with different service providers, service providers say, “Wait a minute.” We’re seeing it in everything from ride-hailing to food services, and how these relationships play out will help determine which AI investments will pay off.
If you make a purchase through links in our articles, we may earn a small commission. This does not affect editorial independence.
