AIToday
AI Regulation & PolicyAI Business & IndustryTechCrunch AIPublished: Nov 19, 2026, 04:00 JST

Citibank puts AI-agent losses on customers in Online User Agreement update

Send on LINE
Citibank puts AI-agent losses on customers in Online User Agreement update

3 Key Points

  1. What happened

    Citibank changed its Online User Agreement this week so customers who grant third-party AI tools access to their accounts are responsible for what those agents do — even if the tool leaks credentials or is hacked. JPMorgan, Bank of America and PNC announced similar policies.

  2. Why it matters

    Customers using AI assistants for spending analysis, advice or bill payments now carry the loss if a transaction goes wrong or hackers exploit a leaked credential, shifting the risk away from the bank.

WHO IT HITSRetail banking customers who hand account credentials to personal-finance AI tools now bear the financial risk of unauthorized transactions or breaches, while the banks themselves are shielded from reimbursing losses.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The move by Citibank follows similar policies from JPMorgan and Bank of America last month, and PNC changed its terms today, suggesting a coordinated shift among large banks rather than a one-off. Citibank's update goes further than assigning responsibility for an agent's transactions: it also places liability on the customer if the AI tool leaks credentials or is hacked, a scenario banks have little ability to monitor or reverse.

The timing reflects a highlighted tension. AI startups are looking for revenue and see personal finance as an attractive market, but banks view these tools as both an operational risk and a competitive threat. Markets briefly shuddered at the prospect of AI tools helping consumers avoid fees or low-interest deposit accounts that pad bank margins. Citibank's terms note that AI agents the bank itself provides can be used without additional liability, drawing a line between in-house and third-party tools.

Anthropic's read-only personal finance tool, launched earlier this month, illustrates one workaround: it cannot move, manage, or invest money. The banks' reaction is part of a broader pattern in which service providers push back after AI labs promise to transform how people interact with them, from car-hailing to food service.

FAQ
Does this apply if my AI assistant gets hacked?
Yes. Citibank's updated terms say the customer is responsible for anything hackers do with their account if the AI tool leaks credentials or is hacked.
Which other banks have similar policies?
JPMorgan and Bank of America announced similar policies last month, and PNC changed its terms today. Citizens also emailed customers this week with a similar warning.
Are any AI finance tools designed to avoid this risk?
Anthropic's personal finance tool, launched earlier this month, is billed as read-only and cannot move, manage, or invest money.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleOneTrust Fall '26 Release adds continuous AI inventory