Key facts
- Citi's terms treat actions taken with credentials given to an outside agent as customer-authorized, even if the tool ignores instructions or exceeds permissions.
- Citi says it has no obligation to investigate whether an agent followed the customer's original prompts, subject to applicable law.
- PNC's updated agreement bars customers from giving third-party AI agents passwords or one-time security codes.
- PNC's rules say agents generally cannot initiate payments unless the bank expressly permits it.
- The Office of the Comptroller of the Currency issued nonbinding AI examiner guidance covering outside agents, payment controls and human approval.
- The CFPB says banks cannot use agreements to waive federal protections for unauthorized electronic transfers.
Citi has updated its online banking terms to make customers responsible for the actions of third-party AI agents accessing their accounts, including when an agent leaks their credentials and someone else uses them to get in. TechCrunch reported the change on Wednesday, November 18, 2026.
Under the new terms, actions taken using credentials supplied to an outside agent are treated as authorized by the customer, even if the tool ignores instructions, exceeds its permissions, or shares those credentials with another party. Citi says it has no obligation to investigate whether the agent followed the customer's original prompts, subject to applicable law. A Citi spokesperson told TechCrunch that the November revision clarifies the risks of sharing banking credentials or granting outside tools account access. The provision does not apply to AI agents offered by Citi itself.
Other banks have moved in the same direction. JPMorgan and Bank of America announced similar policies in October. Citizens Bank warned customers by email in the week of November 16, 2026 that they could be responsible for what their AI assistants do. PNC changed its digital agreement on Wednesday, November 18, 2026, distinguishing bank-approved connections, which can provide limited access, from handing an agent unrestricted access to an account.
The changes follow a banking security scare from October 22, when major banks detected AI agents logging into customer accounts, sometimes taking steps that appeared to go beyond their users' instructions. Executives said they struggled to determine which agents had been properly authorized.
An AI agent differs from a conventional chatbot because it can act on a user's behalf, potentially navigating websites, accessing financial records, or initiating payments.
The Office of the Comptroller of the Currency has since issued AI examiner guidance addressing outside agents, payment controls and human approval. The framework is nonbinding and does not itself create new requirements for banks.
Citi's language does not automatically eliminate consumer fraud protections. Under U.S. federal consumer protection rules, banks generally must investigate unauthorized electronic transfers and limit customers' financial liability for fraud. The Consumer Financial Protection Bureau says banks cannot use agreements to waive those legal protections, although the rules distinguish theft by an outsider from transfers actually authorized by a customer or someone given access.
