World ID wants you to put a cryptographically unique human identity behind your AI agents

World ID is being positioned as a way to attach a cryptographically secure, unique human identity to AI agents, with iris-scan backed tokens aiming to curb Sybil attack-style floods of requests. Ars Technica reports that World now claims nearly 18 million unique humans have verified their identities at around 1,000 physical orbs worldwide, and Agent Kit would let those users tie their confirmed identity to AI agents operating on their behalf.

The system sits on the x402 protocol, which was built with CloudFlare and Coinbase support, and some sites have already used micropayments via that protocol to verify authenticity. World says about 18,000 new users have confirmed their identities in the last week, but adoption hinges on the availability of a compelling app that requires such biometric verification.

The piece notes that WorldCoin still exists, though its value had fallen from early 2024 peaks, and frames World ID as a potential framework for authentic AI agent interactions across the web, including access to limited resources like reservations or tickets. According to World, the aim is to reduce automated astroturfing and dogpiling while enabling genuine human oversight of agent activity.