- World ID for Zoom will put a Human Badge on your video feed to prove you're... human
- Setting this up requires a physical iris scan and a verification selfie
- Verifying human authenticity is easier than identifying deepfakes
Sam Altman's biometric identity verification company, World, has launched a series of tools designed to help businesses distinguish real people from AI deepfakes as one of his other companies – OpenAI – becomes increasingly capable of mimicking humans.
One of the main announcements was the general availability of World ID for Zoom, which gives enterprise customers access to a verification process that can distinguish genuine video from AI content.
Under the newly-launched system, verified participants will receive a "Human Badge" displayed alongside their video feed so that other participants can be more confident that they're speaking with a human.
World ID for Zoom
While the problem World is aiming to solve seems rather dystopian and unlikely, recent Gartner research actually revealed that 36% of senior cybersecurity leaders reported at least one incident involving a deepfake during an video call over the past 12 months.
We also know that Microsoft has been launching numerous updates to address similar concerns in Teams, opening up its platform to enable third-party deepfake detection in the hope of stamping out AI-generated video and audio.
However, proving human authenticity isn't as simple as ticking a few boxes or scanning a passport. Workers will have to visit one of World's physical Orb devices, which scans their iris to create a cryptographically secure biometric reference.
To streamline the process, enterprises can choose to have an Orb installed onsite, though this could likely be a costly option.
A selfie through the mobile app then verifies the human participant against that biometric reference image during Zoom calls.
World says that detecting deepfakes can be especially challenging, "lead[ing] to an arms race between deepfake detection and creation," hence the decision to verify human authenticity instead.
