World launches Agent Kit beta for human-verified AI agents

Identity startup World has released a beta version of Agent Kit, allowing users to link their iris-scan verified World ID to AI agents. The tool aims to help websites distinguish requests from human-directed agents amid rising concerns over AI agent swarms. It builds on iris-scanning technology originally tied to the Worldcoin cryptocurrency.

World, the company behind the Sam Altman-founded Worldcoin launched in 2023, announced the beta of Agent Kit on March 17, 2026. The kit enables users who have verified their identity via iris scans on one of nearly 1,000 physical orbs to attach a cryptographically unique World ID token to their AI agents. Nearly 18 million unique humans have completed this verification process worldwide, with about 18,000 new users in the past week alone. Worldcoin, which offered free tokens for iris scans, continues to exist though its value has fallen below early 2024 peaks, as World pivots focus to World ID tokens stored on phones for secure online identity proofing. Tools like OpenClaw have demonstrated how users can deploy swarms of AI agents for tasks, creating Sybil attack-style overloads akin to DDoS for online services. Agent Kit addresses this by letting websites require AI agents to present a World ID token, confirming a real human directs them. This could limit access to resources such as restaurant reservations, ticket sales, free trials, or bandwidth, while preventing automated abuse in forums and polls. The system relies on the x402 protocol, developed with support from Cloudflare and Coinbase. While some sites use x402 for micropayments to rate-limit agents, World argues that unique World ID tokens cannot be faked by attackers, unlike payments which can be funded in bulk. Adoption remains a hurdle, as widespread use depends on more people undergoing biometric verification amid a chicken-and-egg challenge for online human identity assignment.

Relaterte artikler

A bustling crowd of developers at an AI and blockchain hackathon in Miami Beach, building innovative projects with AI agents and drones.
Bilde generert av AI

AI agents spark startup frenzy at consensus miami hackathon

Rapportert av AI Bilde generert av AI

Nearly 1,000 developers gathered in Miami Beach for the EasyA Hackathon at Consensus Miami 2026, building AI-native projects focused on autonomous payments, drones, and consumer apps. The event highlighted a shift toward practical AI and blockchain applications.

Coinbase has introduced Agentic Wallets, a new infrastructure designed to enable AI agents to handle financial transactions autonomously. Announced on February 11, 2026, the tool addresses limitations in current AI capabilities by allowing agents to spend, earn, and trade without constant human oversight. This development builds on prior Coinbase initiatives in agentic commerce.

Rapportert av AI

Developers are equipping increasingly autonomous AI agents with crypto wallets, enabling them to hold assets, trade tokens, and hire other agents, according to Electric Capital. While the technical infrastructure is advancing, legal frameworks lag behind. At NEARCON 2026, Avichal Garg highlighted the unresolved issues around liability for such non-human entities.

Discord has informed UK users that they may be part of an experiment using the age-assurance vendor Persona for verification, where submitted data is temporarily stored unlike previous promises. This change has raised privacy concerns among users, particularly due to Persona's links to investor Peter Thiel and his surveillance firm Palantir. The update is part of a broader global rollout of mandatory age verification starting in early March.

Rapportert av AI

Illia Polosukhin, co-founder of NEAR, predicts that AI agents will serve as the main users of blockchain technology. He envisions AI acting as the front-end interface for online activities, including crypto, while blockchain operates as the back-end infrastructure. This perspective suggests a shift where humans interact primarily through AI, abstracting complex blockchain elements like wallets and transaction hashes.

Warp has released the source code for its terminal client on GitHub under dual MIT and AGPLv3 licenses. The company is shifting its development model to emphasize human oversight on features and reviews, with AI agents handling code implementation. CEO Zach Lloyd announced the move to foster community contributions through this agentic workflow.

Rapportert av AI

Anthropic has restricted unlimited access to its Claude AI models through third-party agents like OpenClaw, requiring heavy users to pay extra via API keys or usage bundles starting April 4, 2026. The policy shift, announced over the weekend, addresses severe system strain from high-volume agent tools previously covered under $20 monthly subscriptions.

 

 

 

Dette nettstedet bruker informasjonskapsler

Vi bruker informasjonskapsler for analyse for å forbedre nettstedet vårt. Les vår personvernerklæring for mer informasjon.
Avvis