Scientist launches tool to verify humans behind AI shopping agents

World, co-founded by Sam Altman, is dedicated to creating what it calls “human proof” technology — identity verification tools for an Internet that is increasingly overrun by AI-generated content of questionable quality. And the connection has not been lost, as Altman’s other company, OpenAI, has been widely blamed for creating much of this failure (although one could argue that he saw the problem coming when he founded World).

This week, Tools for Humanity (TFH), the startup behind World, released the beta version of its new verification tool — this one designed to support building proxy commerce, the fast-growing practice of using artificial intelligence programs to browse the web and make purchases on a user’s behalf.

More and more consumers are using AI agents to browse websites and buy things for them. This trend promises a certain amount of automated convenience, but it has also raised the specter of new forms of fraud, spam, and other forms of widespread Internet abuse.

On Tuesday, World announced its alleged solution: AgentKit, a software development tool geared toward merchant websites that allows the inclusion of a new verification system that lets those sites verify that a real human is behind an agent’s purchasing decisions.

AgentKit is based on World ID, the backbone of TFH’s verification system. The most secure version of the ID is obtained from scanning the user’s eyes via the World’s Orb device. The Orb converts the iris into a unique, encrypted digital token – the Verified Global Identifier – which can then be used to access the TFH services ecosystem via the company’s World app.

AgentKit allows the user’s global identifier to be integrated into the recently launched payment system known as the x402 protocol. Developed by Coinbase and Cloudflare, x402 is an open, blockchain-based standard to allow automated computer programs to transact with each other directly over the Internet – without human intervention at every step. To use AgentKit, users simply register their AI agents using their Global ID, which then communicates to websites – via the x402 system – that a distinct, verified human approves the agent’s purchasing decisions.

“AgentKit was designed as a complementary extension to the x402 v2 protocol, coordinated by Coinbase,” Tools for Humanity said in a statement. “The integration is designed so that any website already using x402 can enable unique human verification proof alongside (or in place of) micropayments.”

In an interview with TechCrunch, Thiago Sada, chief product officer at TFH, compared the new functionality to delegating a “power of attorney” to an agent. By verifying that an AI program is acting on behalf of a particular user, a website can decide whether or not to trust transactions initiated by those agents, Sada said. “What the Universal Identity Badge tells you is that someone is a real, unique human being,” he said, noting that websites can still choose to block certain users they believe are acting in bad faith.

AgentKit is currently being offered as a beta to developers, with the hope that feedback will improve it over time. Sada also noted that consumers would need to have a verified global ID, derived from an orb scan, to qualify for this type of verification.

It’s a step in time. Major e-commerce sites and financial services have already begun to embrace proxy commerce. In the past year, companies like Amazon and Mastercard have introduced automated purchasing capabilities to their platforms, and Google recently launched its own protocol designed to support this trend. As this field grows, it is clear that the industry will need safeguards to ensure that it remains reliable and stable. It is clear that the world is trying to position itself as the actual provider of this stability.