Citibank, the third-largest commercial bank in the United States, changed its terms of service this week, telling customers who use third-party AI tools for personal finance that anything that goes wrong is their problem — even if they’ve been hacked.
AI agents, from frontier laboratory models to purpose-built personal finance tools, can obtain credentials for your bank account and other online financial portals, with the idea that they can analyze spending and investment decisions, provide advice, and even handle transactions and pay bills.
But financial institutions are not happy about the presence of untrusted AI actors in their systems, and are telling customers that they will not be held liable for reversing an intentional or unintended transaction conducted by automated assistants.
“As third-party AI tools become more widely available, we want customers to understand the risks associated with sharing banking credentials or allowing these tools to access their accounts,” a Citi spokesperson told TechCrunch. “We recently updated our Online User Agreement to provide greater clarity around those risks and to reinforce that customers who choose to grant account access to third-party AI agents are responsible for activities conducted through those agents.”
JPMorgan and Bank of America, the country’s two largest banks, announced similar policies last month. PNC Bank, the eighth largest bank, changed its terms today.
Citi’s update goes beyond telling clients that they are responsible for anything their agents might do; It also informs them that if the AI tool leaks their credentials or is hacked, the customer will be responsible for anything the hackers do to their account. Likewise, banking giant Citizens sent an email to its customers this week, warning that they too are on the hook for anything their AI assistant does, according to an email seen by TechCrunch.
As AI startups look for more revenue, the foray into personal finance represents an attractive opportunity for companies to add value to customers and exploit the influx of fees generated by this sector. But if consumers are already wary of sharing their email or chat histories with an AI agent, bank and brokerage accounts present a new trust challenge.
Anthropic’s approach is pretty clear: the personal finance tool it launched earlier this month is explicitly described as “read-only” and “cannot move, manage, or invest your money.”
Beyond cybersecurity concerns, banks also view efforts to enter their franchises as a competitive threat. Markets briefly shuddered in response to the prospect of artificial intelligence tools that would make it easier for consumers to avoid unnecessary fees or low-interest deposit accounts that bolster banks’ margins. But if AI agents become an important way to interact with financial decisions, banks will struggle to control this trajectory. Citi’s terms state that clients can use the AI agents it provides without worrying about additional liability.
The banks’ reaction forms part of a broader trend among companies of all stripes: AI labs promise that their agents will revolutionize the way people interact with a variety of providers, and then the providers say, “Wait a minute.” We’ve seen this in everything from ride-hailing to food services, and how these relationships will help determine which AI investments will pay off.
When you buy through links in our articles, we may earn a small commission. This does not affect our editorial independence.

