Banks are changing the terms for people who use artificial intelligence agents to manage their finances. Citibank now warns that, if we give a third-party AI access to our account, we’ll bear the consequences of its actions, even if the assistant comes under attack, according to TechCrunch.
What changes when we give access to our account
An AI agent connected to financial information can be really useful. It can analyze our spending, advise us on different investments and, with the necessary permissions, pay bills or carry out certain transactions. The issue is that, to grant this access, we have to hand over our credentials to an external system and the risks of managing identities and access increase substantially.
In response to this situation, Citi’s new terms state that the customer will be fully responsible for transactions carried out by these assistants, even if they were made in error. The bank intends to apply that same standard of responsibility even if the agent is attacked or leaks passwords and someone uses them to access the account. JPMorgan and Bank of America updated their policies with similar changes in October and PNC and Citizens issued warnings this past November.
Security and competition in the banking sector
Beyond security, banks see a threat in how assistants can help us avoid fees. Claude, for example, can analyze our bank accounts, but through read-only access, with no ability to move our money. Citi, for its part, allows us to use its own agents, but not third-party ones, without taking on that extra responsibility for errors.
As purchases managed by AI agents take shape, the most important thing for us is to pay close attention to which permissions we grant. Checking balances and transactions is very different from authorizing payments. And, with these new terms at more and more banks, a mistaken transaction may not be covered by the protection we expected from our bank.