Citibank, the nation's third-largest commercial bank, revised its terms of service this week to inform customers that if they share banking credentials with third-party AI tools for personal finance management, they bear full responsibility for any problems that arise—including breaches and hacks. The policy change, reported by TechCrunch on November 18, reflects mounting concern among major financial institutions about artificial intelligence agents gaining access to customer accounts—and a determination not to shoulder liability for the technology's failures or security risks.

Citibank's move follows similar announcements by JPMorgan and Bank of America, the country's two largest banks, which introduced comparable policies last month. PNC, ranked eighth nationally, updated its terms the same day as Citi. Citizens bank also contacted customers this week with warnings about AI assistant liability, according to correspondence reviewed by TechCrunch. The updated agreements go beyond holding customers accountable for transactions their AI agents execute—whether intentional or accidental—and extend to scenarios where the third-party tool itself is compromised, leaking login information to bad actors who then drain or manipulate the account.

A Citibank spokesperson told TechCrunch that "as third-party AI tools become more widely available, we want customers to understand the risks associated with sharing banking credentials or authorizing those tools to access their accounts," adding that the firm "recently updated our Online User Agreement to provide greater clarity about 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." Citi's revised terms clarify that AI agents provided directly by the bank can be used without customers facing additional liability. Anthropic's personal finance tool, launched earlier this month, is explicitly marketed as "read-only" and "can't move, manage, or invest your money," according to the report.

The policy shift reflects dual anxieties within the banking sector. Financial institutions view AI assistants that analyze spending, recommend investments, and execute bill payments as both a cybersecurity vulnerability and a competitive threat. AI startups see personal finance as an attractive revenue opportunity, tapping into the lucrative fee streams the sector generates while offering customers tools to potentially sidestep unnecessary charges or low-interest deposit products that bolster bank profit margins. Markets reacted with brief volatility to the possibility that AI could help consumers avoid these fees, the report notes. If AI agents become a primary interface for financial decision-making, banks are positioning themselves to control that channel rather than cede it to outside technology firms.

The banking industry's response is part of a wider pattern across service sectors: AI laboratories pledge that their agents will transform how people engage with everything from transportation to food delivery, then the actual service providers push back. These standoffs—visible in car-hailing, dining, and now finance—will shape which AI investments ultimately deliver returns and which fall short. For consumers already hesitant to grant AI access to email or chat histories, bank and brokerage accounts represent an even steeper trust barrier, one that the new liability terms make considerably harder to cross. Institutions willing to build competing AI tools themselves may control the relationship, while those relying solely on disclaimers risk ceding ground to rivals who offer both convenience and protection. The question isn't whether AI will touch finance, but who writes the rules when it does.