World, an identity project cofounded by Sam Altman, has launched AgentKit, a toolkit that lets AI agents carry cryptographic proof they are backed by a unique human via the World ID system. The beta release arrived on Tuesday as websites increasingly need ways to distinguish legitimate automated activity from spam, bot swarms, and abuse in the growing world of AI commerce.
The core problem AgentKit addresses is deceptively simple. One person could run thousands of agents that all pay small fees. Without human verification, a bad actor deploying large coordinated agent swarms could overwhelm booking systems, drain free trial allocations, scalp ticket releases, or manipulate online forums. By integrating with Coinbase and Cloudflare's x402 protocol for stablecoin micropayments, AgentKit aims to make AI agents verifiable economic participants rather than suspicious automated traffic.
The economics of this problem are becoming urgent. McKinsey estimates agentic commerce could generate $3 trillion to $5 trillion globally by 2030, while Bain has projected AI agents could account for as much as 25% of US ecommerce sales by the end of the decade. As these autonomous systems transition from toy demonstrations to real transactions, platforms can no longer simply block all automated traffic.
AgentKit works by linking multiple agents to a single verified human identity. AgentKit links multiple agents to a single verified person using zero-knowledge proofs and, for now, Orb-based biometrics, enabling platforms to cap usage per human and positioning World as a foundational identity layer for an AI-driven web. A developer registers their agent with their World ID; when that agent accesses an x402-enabled website, the site can request proof of human backing alongside or instead of micropayments.
According to Erik Reppel, head of engineering at Coinbase Developer Platform and founder of x402, "Payments are the 'how' of agentic commerce, but identity is the 'who."" The approach differs from existing rate-limiting solutions because micropayments alone fail to prevent determined attackers. A research engineer at the World Foundation noted, "One person could run thousands of agents that all pay small fees. Proof of Human addresses this gap."
World's installed base provides immediate momentum. World reports that its network now includes nearly 18 million verified humans across more than 160 countries. Unlike systems requiring phone numbers, credit cards, or other identity proxies that bad actors can acquire at scale, biometric identity theoretically imposes a hard constraint: each iris is unique and cannot be duplicated.
However, biometric approaches carry substantial trade-offs that warrant scrutiny. Digital rights groups such as the Electronic Frontier Foundation have warned about risks if identifiers are compromised or centralised. Unlike a driver's licence or a passport, you cannot replace your iris if the data is compromised. World emphasises encryption and decentralisation, claiming it uses zero-knowledge proofs so platforms verify uniqueness without collecting personal data. Yet regulators argue that cryptographic codes can still identify people even if the original data is deleted, which under GDPR would count as personal data.
Adoption challenges remain concrete. Countries including Brazil, Colombia, Germany, Hong Kong, India, Kenya, Portugal, Spain, and South Korea are investigating the solution for privacy violations or have already banned the Orb or Worldcoin. The system requires non-standard iris scanner hardware, adding integration costs compared to identity verification schemes using mobile device native sensors.
The chicken-and-egg problem remains. AgentKit only becomes valuable if major platforms adopt x402 and require World ID verification. Yet convincing users to undergo iris scanning—a one-time biometric commitment—depends on applications that genuinely require it. Since the start of 2026, the number of agents using the ERC-8004 standard across blockchain networks has grown from 337 to nearly 130,000. That explosive growth in agent activity creates real demand for trust infrastructure, but whether World can become the standard remains far from settled. The technology addresses a genuine problem, yet reasonable people can question whether the privacy costs of permanent biometric identity justify the benefits for fraud prevention in digital commerce.
The system uses zero-knowledge proofs so platforms can verify that an agent represents a real person without collecting or storing personal data, a design World claims is required for scaling identity in an AI-driven web. For more information on how AgentKit works, developers can access the official integration documentation.