Citi updates online user agreement making customers liable for actions by third-party AI agents, including if the tools are hacked or leak credentials
Edited by Elliot Marsh, Editor · written with AI from the linked sources (how)
Citibank changed its Online User Agreement this week so customers who give third-party AI agents account access are responsible for all activity through those agents, including losses if the AI tool leaks credentials or is hacked. Citizens emailed customers with similar warnings, and PNC changed its terms the same day. JPMorgan and Bank of America announced similar policies last month. Citi's terms exempt AI agents provided by Citi itself.
Entities: Citibank, JPMorgan, Bank of America, PNC, Citizens, Anthropic
0 primary
What happened
Citibank changed its Online User Agreement this week. Customers who give third-party AI agents access to their accounts are now responsible for all activity through those agents, including losses if the AI tool leaks credentials or is hacked. Citizens emailed customers with similar warnings and PNC changed its terms the same day. JPMorgan and Bank of America announced similar policies last month. Citi's terms exempt AI agents that Citi itself provides. The agreement text is not linked, and no figures on losses or customer numbers are given.
Why it matters
Anyone who lets an outside AI tool log in to a bank account now carries the full risk if something goes wrong, with no protection from the bank. That affects consumers and small firms, and it raises the cost of building AI finance products, because startups depend on customers being willing to take that risk. Citi's carve-out for its own agents suggests banks want to control this channel rather than ban it. Several large US banks now share the same stance, so it looks like a sector norm rather than a one-off. How far the terms would hold up under existing consumer protection rules is untested and uncertain.
What is noise
The claim that this will decide which AI agent investments pay off, and the framing of a broad service-provider backlash against AI agents, is speculation beyond what one set of terms changes. Banks have long told customers not to share credentials with third parties, so this is partly a restatement of existing risk allocation, and JPMorgan and Bank of America were already known. The motive of protecting fee margins is the writer's reading, not something the banks stated.
Watch next
- 01Whether regulators such as the CFPB or state authorities say anything on liability for customer-authorised AI agents, or whether any dispute tests these terms in practice.
- 02Whether other large banks (Wells Fargo, Capital One, and UK or EU banks) adopt similar terms, and whether banks launch their own agents or approved-partner access schemes.
- 03How AI agent and fintech providers respond, for example with security guarantees, insurance, or limited-permission access that avoids sharing passwords.
Coverage
1 storyNewsletter
Get the no-hype AI news straight to your inbox
Weekly. No spam, unsubscribe any time. Privacy
More adoption signals
Full feed →- Meta releases Muse Spark Open, an open-weight ~816B-parameter multimodal mixture-of-experts model, with Hugging Face support15 Oct 202680
- Google releases Gemini 4 Neon, priced at $3/$15 per million tokens with a 1M-token context window, rolling out across API, Antigravity and the Gemini app18 Nov 202679
- Google DeepMind launches AlphaGenome Atlas, a free public database of predicted effects for 9 billion possible human genome variants8 Sept 202679
- Apple ships Gemini-powered Siri beta with iOS 27, excluding EU and China at launch15 Sept 202678