Citi updates online terms to make customers liable for third-party AI agent actions

For the general counsel at a regional US bank

Citi updated its Online User Agreement in November to make customers responsible for actions taken by third-party AI agents they give account access, according to a Citi spokesperson quoted by TechCrunch.31

Citi's spokesperson told TechCrunch the bank wants customers to understand the risks of sharing banking credentials or authorizing AI tools to access their accounts.3 Yahoo Finance reported the change follows an Oct. 22 scare in which major banks detected AI agents logging into customer accounts, sometimes beyond their users' instructions, and executives said they struggled to tell which agents were authorized.1

Under the new terms, actions taken with credentials supplied to an outside agent are treated as authorized by the customer, even if the tool ignores instructions, exceeds its permissions or shares the credentials with another party, according to Yahoo Finance's reading of the terms.1

Citi says it has no obligation to investigate whether the agent followed the customer's original prompts, subject to applicable law, and the provision does not apply to AI agents Citi itself provides.13

TechCrunch reported that JPMorgan and Bank of America announced similar policies last month, that Citizens emailed customers this week with a similar warning, according to an email TechCrunch saw, and that PNC changed its terms today.3

PNC's updated agreement lets customers use certain bank-approved connection methods but bars giving AI agents passwords or one-time security codes, and agents generally cannot initiate payments unless PNC expressly permits it, according to Yahoo Finance.1

Why it matters

Banks are shifting agent-related loss risk to customers by contract, but federal rules on unauthorized transfers still apply and the Consumer Financial Protection Bureau says agreements cannot waive them. Executives should expect disputes over whether agent-driven losses count as unauthorized, and should weigh technical controls like PNC's against contract language alone.

From the brief

Our take. Citi now treats a hacked agent's transfers as customer-authorized; JPMorgan and Bank of America have similar policies. Federal rules say contracts cannot waive fraud protection. Your terms must pick a side before your first agent dispute does.

Action. Legal and compliance: Map digital-banking terms on third-party agent credentials against unauthorized-transfer investigation duties, and decide whether to ban password and one-time-code sharing.

What our records say

Citigroup treats AI as the instrument for completing Jane Fraser's multi-year overhaul rather than as a separate program.1 It reports results on the quarterly earnings deck.1 Fraser set out a four-bucket framework for AI returns: business strategies, end-to-end productivity, defensive capabilities (cyber, fraud, AML) and workforce implications (Apr 14, 2026).2

  1. Citi: AI programme summary (part 1)
  2. Citi: Q1 2026 call: a four-bucket AI framework

From our records

Sources

  1. ReportCiti Says Customers Are Responsible When AI Banking Agents Go Rogue—Even After a Hack Yahoo Finance
  2. Primary sourcehttps://www.coinbase.com/stock/C?utm_source=yahoo_finance&utm_campaign=cx_o_m_w_m_m_yah_gro_dis_yf&utm_medium=display&utm_content=promo%3DYAHOO_1_MONTH Coinbase
  3. Also reported byMore banks says customers using AI agents are on their own — even if they get hacked TechCrunch