World ID matches an Orb scan, a phone selfie, and the live video frame. It can secure internal approvals among enrolled staff, but it cannot verify an unknown caller.
AI fraud is growing into a measurable financial risk. Deloitte estimates, in figures cited by Zoom, that US losses could jump from $12.3 billion in 2023 to $40 billion by 2027. That is a 3.25-fold increase.
A visual badge confirming a real human sits behind a webcam sounds like a possible fix. World's Human Badge for Zoom provides that verification, but only for people who have already enrolled.
The integration shifts the focus from spotting synthetic video to checking a roster of enrolled users. This setup stops fraud when an organization controls both ends of the meeting, such as board votes and fund transfers between verified employees. It cannot stop impersonation fraud from an unknown caller, which is how most social engineering attacks start.
World announced in a post dated 18 October that World ID for Zoom "is now available" to Zoom's enterprise customers globally through the Zoom Marketplace. Zoom and Tools for Humanity, the firm behind World, previewed the integration earlier this year.
October marks its general availability, alongside a sales route through the Zoom ISV Exchange Program. Under this arrangement, Zoom enterprise clients can install Orbs, World's proprietary iris scanning hardware, inside their own offices.
The verification workflow reveals both its strength and its limits. First, a participant links a World ID to a Zoom account. During a call, the participant receives a mobile notification and takes a selfie in the World app.
World and Zoom's Marketplace listing explain that the system verifies three distinct images: the Orb scan signed during the initial Proof of Human registration, the fresh selfie from the user's phone, and the live video feed. When all three match, the Human Badge appears on the participant's video tile for everyone on the call to see.
The hardware iris scan happens only once at an Orb. Subsequent meetings require only the mobile selfie and the live video feed. World notes that the software, built on Zoom's Realtime Media Streams, "analyzes video only, not audio or transcripts".
Because of this design, the badge offers no protection against cloned voices on audio lines. It also offers no protection against unverified participants joining the room.
Authentication runs entirely inside World's app. Its FAQ states that face matching happens on the user's phone. World transmits only a binary yes-or-no verification token to the World ID for Zoom app rather than to Zoom directly. Zoom's release states that no underlying data reaches Zoom or other attendees. It adds that Zoom and enterprise clients receive attestations verifying that a participant is human and matches their registered profile.
Incomplete coverage remains the weak point. Every participant can see the badge, meaning an unverified guest can see who holds credentials. On a five-person call where three participants hold credentials, three video tiles display badges and two do not. Those two unbadged tiles mark where an attacker can hide.
For financial authorizations, effectiveness rests on how many authorized signatories hold a Proof of Human. World claims that more than 18 million people hold one, according to its post on banking and finance.
One important preventative feature has not launched yet. World states that hosts "can soon" turn on a Waiting Room where participants must submit their Proof of Human before entering. Meeting hosts can already prompt attendees for verification during a live session.
Until that Waiting Room rolls out, verification happens entirely after the meeting begins. An impersonator is already inside the call before the badge fails to show up. World's design partners include Bain Capital, OpenExchange, Outtake, Pantera Capital, and VanEck Funds.
World simplified corporate onboarding by anchoring hardware distribution to the workplace. The company's solutions page notes that World ships an Orb to the client before launch. A World representative visits the office to install the device, with setup usually taking under 60 minutes.
After completing the one-time anonymous scan at the office Orb, employees only need Zoom and the World ID mobile app installed. Administrators can also set badge expiration windows before users must re-authenticate, turning verification frequency into a configurable company policy.
Direct employer deployment sidesteps consumer resistance around iris biometrics. Instead of trying to register the general public, an organization only needs to enroll a defined list of signatories, board directors, and treasury staff. The legal responsibility moves to internal human resources and data protection teams, who must establish a lawful basis for processing biometric data across the EU and UK.
For an organization already managing strict compliance workflows, registering twenty corporate signatories is straightforward compared to broader public adoption campaigns.
Automated deepfake detection represents an alternative path, but it operates under different trade-offs. World argues that frame-by-frame predictive analysis "can lead to an arms race" between synthetic generation and detection engines. However, predictive scanners evaluate unknown participants without prior setup, giving them an advantage when screening an unfamiliar client on a first call.
Authorizing high-value wire transfers requires definitive identity verification rather than statistical probability. The superior model depends on whether an enterprise faces greater risk from known insiders or external attackers.
These dynamics point to a focused deployment. The Human Badge fits structured Zoom approvals like wire transfers and binding corporate resolutions. It delivers value when reserved for authorized personnel who can execute transactions, supported by an Orb stationed in the office that holds them. Organizations can then make the badge mandatory for transaction sign-offs.
Firms still need out-of-band callback procedures for unverified external parties. Wire desks routinely combine hardware security tokens with telephone callbacks, and the Human Badge adds an extra layer of visual dual control. The friction comes when an authorized signer travels or a new counterparty must use standard manual clearance.
Two checkpoints by 31 March 2027 will determine if the Human Badge functions as a real security control. First, whether World ships its promised pre-meeting Waiting Room. Second, whether any of the five design partners publicly confirms that it blocks financial approval calls lacking the badge. Until organizations mandate the badge to authorize transactions, Zoom has added a visual tag.