World introduces verification tool to confirm humans behind AI shopping agents

World, the identity-focused project co-founded by Sam Altman, is centred on building what it calls "proof of human" technology — verification systems designed for an internet increasingly saturated with AI-generated content of uncertain quality. The connection is notable, given that Altman's other company, OpenAI, has often been criticised for contributing to the very surge of AI-generated material that World now aims to address (though some argue this reflects foresight rather than contradiction).

This week, Tools for Humanity (TFH), the startup behind World, introduced a beta version of a new verification tool to support the rise of agentic commerce — a rapidly growing trend in which AI agents browse websites and make purchases on behalf of users.

As more consumers turn to AI-powered agents to handle online shopping, the convenience is clear, but so are the risks. The rise of automated purchasing systems has raised concerns around fraud, spam, and large-scale abuse across the internet.

In response, World unveiled its proposed solution on Tuesday: AgentKit. This software development kit is for commercial websites and helps you integrate a verification system that confirms whether a real human is ultimately behind an AI agent's purchasing decisions.

AgentKit is built on World ID, the core component of TFH's identity system. The most secure form of World ID is generated through an iris scan using the company's Orb device. The Orb converts a user's eye scan into a unique, encrypted digital code, which becomes their verified World ID. This ID can then be used to access TFH's ecosystem via the World app.

Through AgentKit, a user's World ID can be linked to the x402 protocol, a recently introduced payment system developed by Coinbase and Cloudflare. The x402 protocol is a blockchain-based open standard that enables automated systems to conduct transactions directly with one another online, without requiring human input at every step.

To use AgentKit, individuals register their AI agents with their World ID. This setup allows websites, via the x402 protocol, to verify that a distinct, authenticated human is authorising the AI agent's actions.

"AgentKit is built as a complementary extension to the x402 v2 protocol, in coordination with Coinbase," Tools for Humanity said in a statement. "The integration is designed so that any website already using x402 can enable proof of unique human verification alongside (or instead of) micropayments."

In an interview, TFH Chief Product Officer Tiago Sada compared the concept to granting an AI agent "power of attorney." By confirming that an agent is acting on behalf of a verified individual, websites can determine whether to trust transactions initiated by that agent. "What the World ID badge tells you is that someone is a real and a unique human," Sada said, adding that platforms can still block users they suspect of acting in bad faith.

AgentKit is currently available in beta for developers, with TFH aiming to refine the system based on feedback. Sada also noted that users must have a verified World ID — obtained through an Orb scan — to use the feature.

The timing aligns with broader industry momentum. Major players in e-commerce and finance have already begun adopting agentic commerce models. Last year, companies like Amazon and Mastercard introduced automated purchasing features, while Google recently launched its own protocol to support this shift. As adoption grows, the demand for systems that ensure trust, security, and accountability is rising — a space World is positioning itself to lead.