POLICY · MONEY · #3850
Citi updates terms saying customers using third‑party AI agents bear liability if accounts are accessed
Citibank updated its Online User Agreement to state that customers who grant third‑party AI agents access to their bank accounts are responsible for activities conducted through those agents — including if the agents are hacked or leak credentials. Other major U.S. banks (JPMorgan, Bank of America, PNC, Citizens) have issued similar policies, while some AI tools (e.g., Anthropic’s personal finance product) are explicitly read‑only.
KEY POINTS
- Citibank updated its Online User Agreement to state that customers who grant third‑party AI agents access to their bank accounts are responsible for activities conducted through those agents — including if the agents are hacked or leak credentials.
- banks (JPMorgan, Bank of America, PNC, Citizens) have issued similar policies, while some AI tools (e.g., Anthropic’s personal finance product) are explicitly read‑only.
- This signals a broad banking industry shift to push liability for automated AI‑driven account actions onto customers, shaping how consumer finance AI products can be used and trusted.
WHY IT MATTERS
This signals a broad banking industry shift to push liability for automated AI‑driven account actions onto customers, shaping how consumer finance AI products can be used and trusted.