Meta's Muse, OpenAI's dots, and Instinct are personal AI agents now using World ID for proof-of-human verification across consumer apps and services.
AI & Agents ·
Meta's Muse, OpenAI's dots, and Instinct have entered mainstream use as personal AI agents that execute tasks on behalf of consumers—including sending emails, booking travel, completing forms, and making purchases. Dots, launched three weeks after Muse, can integrate with more than 4,000 apps and operate continuously on cloud-based systems, while Instinct functions as an invite-only text-activated assistant. These agents now face a practical challenge: websites and services need assurance that a real person authorized each action, not just that an AI is operating autonomously.
World ID's Proof of Human standard addresses this verification gap using zero-knowledge proofs, allowing agents to confirm human authorization without exposing personal identity. The distinction matters because the pool of humans is finite—roughly eight billion—while the number of possible AI agents is unlimited. Rate limiting, sales attribution, and purchase allocations all depend on counting the individuals behind agent requests rather than the agents themselves. A single person can control thousands of agents yet remain counted as one human.
The framework is designed to be platform-agnostic and open, enabling agents built by any developer to operate with verified human backing. Real-world friction has already emerged: Amazon blocked Muse from its shopping platform shortly after launch, citing disclosure requirements. The ongoing question remains how broadly websites and services will adopt agent verification standards, and whether existing commerce and service rules will evolve to accommodate human-backed autonomous systems at scale.