Citi, PNC tell customers their AI agents are their problem, hacks included
Citibank rewrote its Online User Agreement this week, putting customers on the hook for anything a third-party AI agent does with their bank account — even if that agent leaks credentials or gets hacked. JPMorgan, Bank of America and PNC rolled out similar terms, while Citizens warned customers by email.
- Original source: techcrunch.com.
- The source published it on Nov 18, 2026 at 5:32 PM.
- No second outlet has reported it yet. Treat it as a single-source claim.
Citibank updated its Online User Agreement this week, telling customers that if they hand their banking credentials to a third-party AI finance tool, the fallout is on them. That covers intentional and accidental transactions by the agent, and it stretches to hackers too: if the AI tool is breached or leaks credentials, Citi says the customer owns whatever happens next. Citi is the third largest commercial bank in the US.
"As third-party AI tools become more widely available, we want customers to understand the risks associated with sharing banking credentials or authorizing those tools to access their accounts," a Citi spokesperson told TechCrunch, saying the update was meant to clarify those risks and reinforce that customers who grant access to AI agents are responsible for what those agents do.
Citi isn't first out of the gate. JPMorgan and Bank of America, the country's two biggest banks, announced similar policies last month, and PNC, the eighth largest, changed its terms the same day Citi did. Citizens also emailed customers this week with the same message — they're on the hook for whatever their AI assistant does, according to an email seen by TechCrunch.
Banks are pushing back as AI companies angle for a slice of personal finance, where agents can be handed the keys to accounts and portals to track spending, weigh investment calls and pay bills. Citi's updated terms carve out an exception for AI agents the bank itself provides, which customers can use without taking on extra liability.
Anthropic's take shows how touchy the ground is: the personal finance tool it launched earlier this month is billed as "read-only" and "can't move, manage, or invest your money." Beyond security worries, banks read the agent push as a competitive threat — markets briefly wobbled at the idea that AI tools could steer consumers away from unnecessary fees and low-interest deposit accounts that pad bank margins.
The standoff is a familiar pattern: AI labs pitch agents as the future of dealing with service providers, and those providers slam the brakes. It has already played out in car-hailing and food delivery, and TechCrunch notes how these fights resolve will shape which AI bets actually pay off.
TopicsCitibank · JPMorgan · Bank of America · PNC · Citizens · Anthropic · TechCrunch
Source: techcrunch.com
Written by AI from verified reporting.
More stories
- Tech & AI cnbc.com Caught by @joko.wijaya06 2 viewsNvidia interim CEO Kress touts $277B in buybacks, 75% growth target
- Tech & AI x.com 3 viewsKobeissi Letter Slams AI Bubble Doom Talk After Jensen Huang Event
- Tech & AI cryptobriefing.com 3 viewsIntel and Microsoft Push Xeon 6 Deeper Into Azure's Sovereign Cloud
- Tech & AI breakingthenews.net 2 viewsAltman says OpenAI will IPO 'at some point,' warns race dynamic endangers safety

