World, the identity startup formerly known for its WorldCoin cryptocurrency project, has launched Agent Kit beta — a new solution designed to distinguish human-directed AI agents from automated bot swarms that could overwhelm online services.
The problem World is trying to solve has become increasingly urgent. Tools like OpenClaw have demonstrated how tech-savvy users can deploy cadres of AI agents to automate tasks at scale. While convenient for individual users, this creates what amounts to a DDOS-level threat for websites and services suddenly facing thousands of AI-generated requests — a classic Sybil attack scenario where one human appears as many fake identities.
Agent Kit provides a cryptographic verification layer. Website operators can now require visitors to prove a real human is directing any AI agents making requests on their behalf. The system leverages World ID, which stores a cryptographically secure, unique online identity token on users' phones. This identity is tied to the same iris-scanning technology that powered WorldCoin's original 2023 launch.
The timing is notable. WorldCoin, founded by Sam Altman in 2023, initially offered free cryptocurrency to anyone who scanned their iris at one of the company's physical "orb" terminals. While WorldCoin still exists, its token value has fallen significantly from early 2024 peaks. The company has pivoted toward World ID as its primary focus, positioning proof-of-human technology as essential infrastructure for an AI-dominated internet.
The beta launch represents World's attempt to become the standard for human verification in an era of increasingly autonomous AI systems. As AI agents become capable of performing tasks on users' behalf at scale, distinguishing genuine human intent from automated scripts could become a critical infrastructure need for the web.