World, the digital identity project co-founded by OpenAI CEO Sam Altman, has launched a verification package for banks and financial firms as it looks to help businesses identify real people, reduce deepfake fraud and verify AI agents acting on behalf of users.
Announced at the Money20/20 conference, the package combines the World ID app, World ID for Zoom and World ID for Agents. The company is targeting fraud risks that arise when criminals use AI-generated faces, voices or identities to impersonate executives, customers and business partners.
World named Bain Capital, VanEck Funds, OpenExchange and Outtake as design partners. However, it did not identify any bank that has already adopted the new package.
How World ID works on Zoom
World ID for Zoom allows meeting participants to verify that a real, previously verified person is present during a video call. Verified participants can display a badge to help others distinguish them from potential deepfakes.
The feature is designed for sensitive conversations, including financial negotiations and business meetings where impersonation could lead to fraudulent payments or the disclosure of confidential information.
Zoom first introduced the integration with Tools for Humanity, the company behind World, in April 2026. The service is now generally available after a limited beta period. World’s verification system relies on credentials created through its identity network. Its highest level of verification, known as Proof of Human, involves the Orb device, which scans a person’s iris to establish a unique identity credential.
World says its cryptographic system allows users to prove they are verified humans without revealing their underlying World ID to third-party services. However, verification does not automatically guarantee that every interaction is safe or that a verified person is acting honestly.
World extends verification to AI agents
The new package also includes World ID for Agents, a tool that allows AI agents to present credentials associated with a verified human.
As businesses increasingly use AI systems to perform tasks and interact with online services, companies face a new question: how can they establish whether an automated action is authorised by a real person?
World’s approach allows a person to delegate certain identity proofs to an AI agent, which can then present those credentials to a third-party platform. The aim is to help services distinguish agents operating with a verified person’s authorisation from systems that cannot demonstrate that connection.For financial firms, this could be relevant to customer onboarding, employee workflows and transactions that require confirmation of who is giving instructions.
Deepfake fraud remains a major concern
The launch comes as businesses face increasing risks from AI-generated impersonation. In a widely reported 2024 case, an employee at engineering company Arup was deceived into transferring approximately $25 million after joining a video conference featuring AI-generated representations of company colleagues.
World is betting that identity verification can provide an additional layer of protection beyond systems that attempt to detect manipulated video or audio.
The technology still faces questions around privacy, accessibility and adoption. Users must be willing and able to complete the relevant verification process, while businesses must decide when proof of humanity is necessary and how to handle people who cannot or do not want to use biometric verification.
For now, World has announced a commercial offering and named design partners, rather than confirming widespread deployment across banks. Its next challenge will be demonstrating that the system can fit into financial institutions’ security and compliance processes while earning the trust of the people expected to use it.