Summary: Zoom has partnered with World, Sam Altman’s biometric identity company, to allow meeting participants to verify that they are human using World’s Deep Face technology, which cross-references scanned iris biometric profiles with live video to display a “Verified Human” badge. The feature is in response to deepfake fraud that cost companies more than $200 million in the first quarter of 2025 alone, including a $25 million loss at engineering firm Arup, although World’s iris-scanning Orb system faces ongoing regulatory action in Spain, Germany, the Philippines and several other countries.
Zoom has partnered with World, the biometric identity company co-founded by Sam Altman, to allow meeting participants to prove that they are real humans and not AI-generated deepfakes. The integration uses World’s Deep Face technology to compare a participant’s live video stream to their iris-scanned biometric profile and displays a “Verified Human” next to your name when the match is successful. Hosts can enable a Deep Face waiting room that requires verification before someone joins, and participants can request that someone verify themselves during the call.
The feature addresses a threat that has gone from theoretical to costly. In early 2024, engineering firm Arup lost $25 million after an employee in Hong Kong authorized a series of bank transfers during a video call in which all other participants turned out to be an AI-generated fake of their colleagues, including the company’s CFO. A similar attack hit a multinational company in Singapore in 2025. Industry-wide, deepfake fraud exceeded $200 million in losses in the first quarter of 2025 alone, and the average loss per corporate incident now exceeds $500,000.
How verification works
World’s Deep Face takes a three-pronged approach. It cross-references a signed image captured during the user’s original registration via the World’s Orb device, a spherical biometric scanner that photographs iris patterns, with a real-time facial scan from the user’s phone or computer and a live video frame visible to other meeting participants. The verification is only successful when all three entries match. The process runs locally on the participant’s device, and World says no personal data leaves the phone.
The 💜 of EU technology
The latest rumors from the EU tech scene, a story from our wise founder Boris and some questionable AI art. It’s free, every week, in your inbox. Register now!
This is architecturally different from the deepfake detection tools already available on the Zoom marketplace. Pindrop, Reality Defender and Resemble AI products analyze video frames for telltale signs of AI manipulation, flagging synthetic media in real time. Both Zoom and World said that because video generation models are improving rapidly, those frame-by-frame detection methods are becoming increasingly increasingly unreliable. Deep Face completely avoids the detection problem by verifying the person’s identity with a biometric record instead of trying to determine whether the pixels on the screen were generated by software.
The downside is that Deep Face requires participants to have a World ID, meaning they must have visited one of World’s physical Orb devices to have their iris scanned. The network currently has around 18 million verified users in 160 countries and approximately 1,500 active Orbs. That’s a small fraction of Zoom’s user base, limiting the feature’s immediate usefulness. For most meetings, existing framework analysis tools will remain the practical choice. Deep Face is designed for high-risk calls where identity certainty justifies the friction of requiring biometric pre-registration.
The business case
Zoom spokesperson Travis Isaman described the integration as part of the “Open ecosystem approach, giving customers more ways to build trust in their workflows based on what matters most for their use case.” The framing is deliberate. Zoom does not endorse World ID as its default identity layer; it offers it as one option among several in a market that already includes multiple identity verification and deepfake detection tools.
For Zoom, the partnership is defensive. The company’s revenue reached $4.67 billion in fiscal 2025, growing modestly at 3%, and its strategic challenge is to remain the default platform for business communication as competitors add AI Features in all areas. Zoom has responded with AI avatars, an AI-powered office suite, and cross-app AI annotators. Adding human verification addresses a different vector: making Zoom the platform businesses trust for confidential conversations. In a market where a single deepfake call can cost $25 million, that trust has measurable commercial value.
For World, the Zoom integration is a distribution win. The company, which rebranded from Worldcoin in 2024, has struggled to move beyond cryptocurrency-adjacent early adopters. Its partnerships with Visa, Tinder, Razer, and Coinbase have expanded the contexts in which a World ID is useful, but none of those integrations create the kind of immediate, visceral demand that a corporate security use case generates. If a company’s treasury team requires World ID verification for any video call that involves wire transfer authorization, that creates institutional adoption that individual consumer associations do not require.
The privacy issue
The World’s Orb-based identity system has faced sustained regulatory scrutiny. Spain’s data protection authority issued a formal warning in February 2026 citing GDPR violations and insufficient data protection assessments. Germany’s Bavarian data regulator ordered the deletion of iris data in December 2024. The Philippines issued a cease-and-desist order in October 2025 to obtain consent through financial incentives. Investigations or suspensions have occurred in Argentina, Kenya, Hong Kong and Indonesia.
He governance frameworks Emerging factors around biometric AI in 2026, including the EU AI Law’s high-risk classification for biometric identification systems, add further complexity. World maintains that its zero-knowledge proof architecture means that verification is performed without exposing personal data, and that iris images are encrypted and stored solely on the user’s device. Critics argue that the recruitment process itself, which requires a physical visit to an Orb to scan its eyes, creates risks that privacy-preserving cryptography does not fully address, particularly when recruitment has disproportionately targeted low-income communities.
For companies evaluating Zoom integration, the calculation is whether the security benefit of biometric human verification outweighs the regulatory and reputational risk of requiring employees or counterparties to register with a company that has been sanctioned by multiple data protection authorities. That calculation will differ by jurisdiction and industry. A Wall Street trading desk pursuing a $100 million Zoom deal may decide the risk is worth it. A European public sector organization almost certainly will not.
What does this mean?
The Zoom-World partnership is an indicator of how far the deepfake threat has advanced. Two years ago, the Arup incident was treated as an extraordinary outlier. Today, deepfake fraud is a billion-dollar category, while AI-generated videos are sophisticated enough to defeat frame analysis detection, and the question of whether the person in a video call is real has become a legitimate enterprise security concern.
The solution Zoom and World are proposing, biometric identity verification anchored to iris scans, works technically but introduces its own set of complications around privacy, regulatory compliance, and the barrier to adoption that Orb’s physical record creates. It’s a feature for specific, high-value use cases rather than a default setting for every Monday morning. But the fact that Zoom thinks it’s worth integrating tells you something about where the technological landscape is heading: toward a future where proving you’re human is no longer something you can take for granted, even when you’re looking someone in the eye.
