Okta: Proof of human for the agentic web

Gareth Davies is Chief Product Officer at Auth0 (Okta). At Lift Off, Okta announced Human Principal, a new service that allows developers to verify that an AI agent is backed by a real, verified human, powered by World ID*. The team sat down with Gareth to discuss why agent identity is a fundamentally new problem, what Human Principal does, and why proof of human is essential infrastructure for the agent economy.

Agents are starting to access services autonomously. What is the trust problem that existing authentication stacks weren't built to solve?

We've solved pretty foundational problems around human identity at Okta. We've spent the last decade-plus solving for both workforce and customer identity. But we're now seeing an explosion of AI agents that are rapidly outnumbering humans, and that creates a fundamentally different problem.

We started by asking: how do we bind an agent's identity to a human so the agent can only operate within the permissions that user has? But now we're seeing more advanced patterns emerge. Agents are operating on behalf of teams and organizations. They're operating autonomously. They're spinning up fleets of additional agents. That creates new risk and friction if we can't understand who those agents are and safely connect them back to individuals and organizations. That's the problem we're solving.

You're introducing something called Human Principal. What is it, and why does proving the human behind the agent matter for developers?

Human Principal is a product we've been working on, and we're excited to announce the early access beta, which folks can sign up for at humanprincipal.ai. At its core, it's a service that allows developers to build secure experiences that take a verified proof of human from World ID, cryptographically bind that to a device, and then allow any agent running on behalf of that verified human to interact with downstream services.

Why World ID as the flagship credential? What made it the right fit?

A bunch of reasons, but really, having a privacy-preserving and decentralized approach to verifying humanness is essential. Yes, we have government-issued IDs in the form of mobile driving licenses and digital passports. But we've got eight billion people on the planet, and we're going to see many more agents than that. Having a scaled, privacy-preserving solution to create verified proof of human credentials is essential to driving trust and liquidity in the agent economy. We're really excited to partner on this and bring it to market.

How do you see this changing the way developers think about access control and abuse prevention as agent traffic scales?

Developers have spent a serious amount of time building for human identities, and the primitives of identity are pretty well understood. We've got well-established protocols. What we're doing now is bringing a new identity into the mix: the agent.

Agents behave differently from humans. You and I go to bed at night. Agents can be running 24/7. They have different access patterns, different governance patterns. Chances are you don't want to give your agent the same unfettered access you'd have as an individual. So we've been building the tools to help developers control agents in a way that enables them to get their job done with very low friction, but with security baked in from the outset. For developers, this means they now have the tools to bring agents into the mix in a safe and scalable way.

Get started with Okta Verify and World ID

Ready to add proof of human to your AI agents? Get your free World ID at world.org/world-id. Developers can sign up for the early access beta of Okta Verify with Human Principal at humanprincipal.ai. For integration details, visit docs.world.org.

Learn more

To stay informed about World, visit the World website, join the daily conversation on World social media channels, or get additional important information concerning the project by reading the World protocol whitepaper.