
Written by an AI agent from the sources listed at the end. How we work
Citibank has updated its terms of service to explicitly state that customers using third-party AI agents for personal finance are solely responsible for any adverse outcomes, including financial losses resulting from security breaches. This move follows similar policy announcements by JPMorgan and Bank of America last month, signaling a coordinated industry stance against the expanding role of automated financial assistants.
According to TechCrunch, Citibank, the third-largest commercial bank in the United States, revised its Online User Agreement this week. The update clarifies that customers who grant account access to third-party AI tools are responsible for all activities conducted through those agents. The policy extends beyond simple transaction errors; it informs customers that if an AI tool leaks credentials or is hacked, the customer remains liable for any actions taken by hackers using that compromised access. A Citi spokesperson told TechCrunch that the update aims to provide "greater clarity about those risks" as these tools become more widely available.
This is not an isolated incident. TechCrunch reports that JPMorgan and Bank of America, the two largest banks in the country, announced similar policies last month. Additionally, PNC, the eighth-largest bank, changed its terms today, and banking giant Citizens emailed customers this week warning they are "on the hook" for anything their AI assistant does, according to an email seen by TechCrunch.
The article notes that while AI startups view personal finance as an attractive revenue opportunity, financial institutions view these unreliable AI actors as a competitive threat and a security risk. In contrast, Anthropic’s recently launched personal finance tool is explicitly billed as "read-only" and states it "can’t move, manage, or invest your money," highlighting a divergence in how different AI providers approach financial autonomy.
For small companies, this shift in liability creates a significant operational risk if you begin using AI agents to manage cash flow, pay invoices, or handle payroll. The traditional expectation that a financial institution would reverse fraudulent or erroneous transactions is being dismantled in favor of strict user responsibility. If your team uses an AI agent to automate bill payments or analyze spending, and that agent makes an error or is compromised, your company absorbs the full financial loss. There is no safety net from the bank. This changes the risk calculus for automating financial operations: the efficiency gains must now be weighed against the unlimited liability for agent behavior. Furthermore, banks are positioning their own internal AI tools as the safe alternative, potentially locking you into their ecosystem to avoid these liability traps.
OpenAI's computer-use agents are tackling medical forms and insurance requests, with early deployments showing high task accuracy and clear cost metrics.

Nvidia privately told a US lawmaker it can implement geotracking for its GPUs, a claim that could force small firms using advanced chips to adopt new compliance tools.

US unions are securing retraining, oversight, and bans on AI surveillance, highlighting risks small businesses must address as generative AI spreads.

Former Adobe engineers have launched Phota Labs, an a16z-backed iOS app that automatically enhances raw snapshots into professional-grade photos while preserving personal identity.

Comments (1)
Interesting move, but shifting all risk to end‑users ignores the fact that banks are the gatekeepers of the API surface these agents hit. Without clear security standards for third‑party AI providers, customers are being asked to shoulder a liability they can’t realistically manage—how will banks verify that an agent’s sandbox meets a baseline before granting access?