Citi makes customers responsible for what third party AI agents do with their accounts

Nov 19, 2026
JPMorgan, Bank of America and PNC have set similar terms as banks push back on AI agents with account access.

Citibank has changed its terms of service to tell customers who use third party AI tools for personal finance that they are responsible for anything those tools do with their accounts, including when the tool leaks their credentials or is hacked.

"We recently updated our Online User Agreement to provide greater clarity about those risks and to reinforce that customers who choose to grant account access to third-party AI agents are responsible for activities conducted through those agents," a Citi spokesperson said. Citi's terms note that customers can use the AI agents the bank itself provides without taking on additional liability.

The move follows similar policies from JPMorgan Chase and Bank of America last month. PNC Bank changed its terms on 18 November, and Citizens emailed customers this week warning that they are on the hook for anything their AI assistant does. Anthropic's personal finance tool, launched earlier this month, is billed as "read-only" and "can't move, manage, or invest your money," according to TechCrunch.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like Finance AI Weekly or Daily AI Brief.

Also, consider following us on social media:

Free newsletter

Finance AI Weekly

Weekly newsletter about AI in finance. Covers AI-driven trading, fintech innovations, and data analytics transforming markets

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more
Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.