World Rolls Out Full-Stack Proof of Human System Amid AI Identity Crisis

World is preparing to unveil its most significant World ID upgrade yet, pitching the new system as a way to prove a person is real online without exposing personal data. In its Lift Off announcement, the company said the updated protocol will introduce “full-stack proof of human,” alongside a soon-to-launch World ID app, while World co-founders Sam Altman and Alex Blania are expected to outline the broader vision and new partnerships during the livestream.

World says the timing reflects a simple reality: bots and AI agents are becoming harder to distinguish from people, while the network of verified humans has already grown to nearly 18 million across 160 countries. At the center of the update is a more advanced version of World ID built around privacy, portability and self-custody.

The company says the new architecture adds features such as one-time-use nullifiers, key rotation, recovery, multi-key support and session management, with the goal of making the system suitable for both everyday users and enterprise deployments. World also says the protocol remains anonymous by design, using zero-knowledge proofs so that a service only learns that a verified human is present, not who that person is.

The consumer side of the launch is aimed at the places where fake accounts and bots have become a daily nuisance: dating, gaming and ticketing. World says World ID is expanding into Tinder globally, with verified users able to show a human badge and, for a limited time, receive five free Boosts.

The company is also introducing Concert Kit, a tool that lets artists reserve tickets for verified humans, and says it will roll out with the Bruno Mars World Tour featuring DJ Pee Wee (aka Anderson .Paak). World’s own framing is blunt: the internet works better when people know they are dealing with a real person, not an automated account.

New Identity Layer for the Internet

World is also pushing the same idea into enterprise security. The company says Zoom will be the first communications platform to integrate its Deep Face feature directly into meetings, using a three-way match between the original Orb verification, a live Face Auth selfie and the current video frame to confirm that the same verified human is present. Docusign is also joining the effort, with World saying proof of humans can help tie approvals back to a verified person in signing workflows.

In the business pitch, World argues that today’s systems verify devices and credentials, but not the human behind them, which leaves room for phishing, theft and session hijacking. The announcement goes further still, extending World ID into the fast-emerging world of AI agents. World says AgentKit now lets verified humans delegate their World ID to agents, creating what it calls human-backed agents with cryptographic proof that a real person stands behind the action.

The company says this is useful for agent delegation, human-in-the-loop approvals and agentic commerce, and it names Browserbase, Exa, Okta and Vercel among the partners and use cases tied to that broader push. World’s message is that the internet is moving from human-driven activity to agent-driven activity, but the trust layer has not kept pace.

What makes the rollout notable is not only the technology, but the positioning. World is not selling proof of human as a replacement for identity documents or a centralized database of personal data. Instead, it is presenting it as a privacy-first signal that can travel across platforms, from a dating app to a concert ticket checkout to a software workflow.

In the company’s view, that makes proof of human a new kind of internet primitive, something users carry with them and services can rely on without storing sensitive details. Whether the market agrees will depend on how easily the system works in practice, but the ambition is clear: World wants to turn human verification into a default layer of online trust.

If the launch lands the way World hopes, the company may have taken a meaningful step toward a more human-first internet, one where bots, deepfakes and agent swarms do not get the final say over who gets access, attention or trust. For now, the pitch is straightforward: in a web increasingly crowded by AI, proof that you are a real person is becoming the product itself.