Muse, dots, Instinct, and the question every agent will be asked
Short summary
As AI agents like Meta's Muse and OpenAI's dots handle tasks on behalf of users, websites need a way to verify a real person is actually behind each agent. World ID's Proof of Human system lets people prove they're unique individuals without revealing their identity, solving the problem that agent numbers are unlimited but humans are finite.
- Amazon already blocked Muse from shopping because the agent didn't disclose it was AI, showing websites will enforce rules about agent transparency and attribution as these tools scale
- Proof of Human matters for practical limits: rate limiting per person, counting unique customers, and preventing one person from gaming purchase allocations by running thousands of agents
- The system uses zero-knowledge proofs so agents can prove human backing without sharing personal data, and companies like Okta, Vercel, and Browserbase are already building it into their platforms
Full summary
AI agents are moving from research labs into everyday use. Meta's Muse handles email, travel bookings, and purchases. OpenAI's dots run constantly in the background and can connect to over 4,000 apps. Instinct works via text message and is spreading fast without requiring an app. These tools put real power in users' hands, but they've already hit friction—Amazon blocked Muse from shopping because it didn't disclose being an AI agent, which violated the site's rules.
The core problem these agents create is simple math: roughly eight billion people exist, but unlimited AI agents can be spawned. Websites need to know if a real human authorized an action, especially for things like rate limiting, sales tracking, and purchase quotas. This is where Proof of Human comes in—a way to verify someone is a real, unique person without revealing their identity. World ID uses zero-knowledge proofs to let people prove they're human and then delegate that proof to their agents. An agent presenting a verified Proof tells a website: a genuine person is behind this request and approved it.
Companies are already building around this. Okta offers Human Principal to verify real humans backing agents. Vercel integrated human-in-the-loop checks into its workflow SDK so people must approve major agent actions before they happen. Exa gives verified humans 100 free API requests monthly across all their agents. Browserbase plans to use World ID to reduce website blocks on verified agents. The pattern is clear: as agents become standard, sites will demand proof of human involvement or block access entirely. World's AgentKit, currently in beta, packages these tools so developers can embed verification into agents from the start.