KOSUNI CRYPTO
Korea's Crypto Pulse, in English
$BTC $ETH $XRP $SOL $DOGE
News
Neutral

World Introduces Human Verification for AI Agents to Confirm Real User Identity

Published October 8, 2026 9:58 AM · 7 views
World Introduces Human Verification for AI Agents to Confirm Real User Identity

World has introduced a solution for verifying whether AI agents are acting on behalf of actual humans, according to an announcement made on August 8. By delegating World ID verification to their agents, users can allow websites or services to confirm that agent actions represent real user requests without receiving personal data. This approach also enables service providers to set limits based on the number of real users rather than the number of agents used by a single person.

Recent AI agents designed as virtual assistants have become popular for handling daily tasks. Meta launched Muse in September, an agent capable of sending emails, booking travel, filling forms, and purchasing products. OpenAI followed with Dots three weeks later, which operates continuously on cloud computers and connects to over 4,000 apps. OpenAI envisions multiple Dots working together as teams to act on behalf of users.

Meanwhile, Instinct, an invite-only service, is seeing rapid user growth and can be used via text messages without a dedicated app on mobile or computer. However, these agents have sometimes clashed with service providers. Amazon blocked Muse shortly after its launch for not disclosing that it was using an AI agent during site navigation.

As agents become more widely adopted, websites and services need to verify both the source of requests and who developed the agent. World ID's human verification allows confirmation without revealing personal information through zero-knowledge proofs. This ensures that even if a user operates thousands of agents, they are counted as one real user.

World is also outlining specific use cases for human verification in limited product sales and high-value transactions. For example, in limited product sales, it verifies whether each agent request represents a unique real user to enforce the 'one person, one agent, one allocation' principle. In high-value purchases or contracts requiring user approval, World ID's human intervention authentication lets users confirm their consent before tasks are completed.

Companies like Octa and Basel have already integrated World ID into their services. Octa offers a beta service called Human Principle to verify if AI agents act on behalf of real people. Basel has incorporated World ID's human intervention authentication into its workflow SDK, allowing developers to add verification steps for user approval at specific task stages.

Other examples include Exa offering 100 free API requests per month to verified users, with all their agents sharing the same limit, and BrowserBase planning to use World ID to reduce agent-based access blocks on websites. World anticipates that as AI agents grow and advanced technologies develop, more platforms will need to confirm if agent requests originate from real humans to avoid being blocked.

Korean Source

This article is an English localization of a Korean-language crypto news report. Original headline: 월드, AI 에이전트 '실제 사람' 확인하는 인간 증명 활용 방안 제시