Citi’s updated user agreement makes customers responsible for activity by third-party AI agents they authorize and, per TechCrunch, for attackers’ actions if the tool is hacked. Citi’s own agents are exempt. PNC and Citizens made similar changes this week.
Reported by: TechCrunch (Nov 18), on Citi’s Online User Agreement. JPMorgan and Bank of America set limits in October.
Your bank’s AI agents are covered. Yours are your problem.
Assuming the bank will reverse losses caused by an agent you authorized.
- Old protections. Fraud rules were written for humans, not delegated agents.
- Fine print. Terms changed by email; finance teams kept connecting tools.
- Hidden access. Nobody knows which agents hold treasury credentials today.
The tell: no one can list which agents can move your company’s money.
When did you last ask which agents can access your bank?
Before year-end close:
- Build the registerList every agent with banking or treasury credentials.
- Test revocationProve you can cut an agent’s access within minutes.
- Reread the termsAsk counsel what each bank now refuses to reverse.
One story a day, and the decision it puts in front of you.
Get Norm's Daily BriefAnalysis and opinion, not investment or legal advice. © 2026 Norm E. Murray. All Rights Reserved.
