Worldcoin: using iris biometrics to prove humanness in the AI era

Executive overview

Distinguishing real humans from bots is becoming impossible with AI-generated deepfakes and fake IDs. Document verification excludes half the world's population and is easily spoofed. Worldcoin uses iris-scanning hardware and zero-knowledge cryptography to issue a privacy-preserving proof of unique humanness — without storing biometric data.

The problem: verifying humans at internet scale

• Bot networks can fake thousands of accounts; no current system reliably detects them
• Document verification is insecure — fake IDs bypass most systems today
• Deepfakes will make document checks even weaker as AI advances
• More than half of adults globally lack a digitally verifiable ID, making document-based systems exclusionary by design

Why biometrics — specifically irises

• Biometrics are the only characteristic that is both globally available and unique per person
• Apple and governments have relied on biometrics for identity for years
• Iris codes are more stable and harder to forge than face or fingerprint data
• A verified iris code can represent uniqueness without revealing identity

How the orb works

• The orb is a purpose-built device that captures and verifies a user's iris in person
• Runs eight neural networks on-device: checks liveness, uniqueness, and anti-spoofing
• Only the iris code leaves the device — all other data is deleted locally
• Open-sourced on GitHub; anyone can inspect the schematics
• Now on its fifth generation; earlier versions physically dispensed a coin with a QR code — scrapped as impractical at global scale

Why crypto and blockchain

• Blockchain is used because it makes privacy provable, not just promised
• Gives users ownership of their identity on the network via tokens
• Creates an open ecosystem — no intermediary permission required to connect
• The goal is decentralisation: the system should be structurally unable to misuse data, not merely trusted not to

AI integration and what comes next

• Expect a wave of co-pilot apps — existing products with embedded assistants rather than AI-native replacements
• The most valuable AI products will solve real, existing problems rather than layering AI onto unproven ideas
• Education will shift from job-focused to passion- and purpose-driven as AI absorbs routine cognitive work
• Disinformation and economic shocks from AI are the near-term risks that deserve more attention than alignment