More banks say customers using AI agents are on their own — even if they get hacked

Impact 7 TechCrunch AI

Summary

  • Citibank updated its terms of service to state that customers using third-party AI agents are fully liable for any financial losses or hacks resulting from that tool.
  • JPMorgan, Bank of America, PNC, and Citizens implemented similar policies, refusing to accept responsibility for transactions or breaches caused by automated assistants.
  • Major banks are pushing back as AI startups attempt to break into personal finance by handling credentials, investments, and bill payments.
  • Anthropic adopted a read-only approach for its recent personal finance tool, avoiding direct management of funds to mitigate liability and trust challenges.

Why it matters

Major banks shifting liability onto customers using AI agents introduces significant financial and operational risk for both end users and AI startups trying to enter the personal finance sector.

Who should care

finance 90