OpenAI Explores a Social Network Using World ID to Curb Fake Accounts

OpenAI would consider creating a social network where registration would require proof of humanity, via World ID and its “Orbs” for eye scan. The idea, reported by several media outlets, is simple on paper: to cut the ground from under the bots and armies of fake accounts, which have become a permanent background noise on X, Instagram or TikTok.

A “humans only” social network, the idea that comes back like a boomerang

Context matters. Sam Altman finds himself at the center of two projects addressing the same issue. OpenAI makes the most widely used AI for the general public. World, led by Tools for Humanity, pushes a digital identity meant to prove that “you are really you,” without necessarily showing your name. Sources even mention a possible mix with Apple’s Face ID as a biometric complement.

This project looks like a direct response to an embarrassing reality. Social platforms need engagement. Yet engagement is manipulated. Bots inflate views, manipulate trends, fuel ghost debates. A “verified humans” network thus becomes both a marketing argument and a technical choice.

This is not the first time OpenAI flirts with the idea of a social platform. The Verge has already reported the existence of an internal prototype described as a social feed like X, with components from ChatGPT. Reuters reiterated this information, specifying that the final format remains uncertain, between a dedicated app and integration into ChatGPT AI.

The strategic logic is clear. A social network is a faucet of fresh content, in real time. It’s also an ideal testing ground for creation, moderation, and recommendation tools. The crux of the matter is the quality of the signal. And bots pollute it.

World ID and the Orbs, promise of cleanliness… and source of tensions

World relies on a physical device, the Orb, to verify a user’s humanity through the iris. Once verified, the user receives a World ID. The project is developed by Tools for Humanity, an organization co-founded by Sam Altman, and it has already been deployed at large scale in several countries.

On the “privacy” side, Tools for Humanity emphasizes a message: the Orb would use data in encrypted temporary memory at the moment of verification, then delete it, with an encrypted copy kept on the user’s device. Said like that, it sounds clean. But this kind of promise depends on implementation details, audits, and trust in governance.

This is where friction appears. Eye scans trigger a visceral reaction in many people. Some see a pragmatic progress. Others see a slippery slope towards an identity imposed by biometrics. World has already been criticized on these topics and has faced recurring regulatory and media questions.

OpenAI’s real bet, beyond hunting bots

A “verified human” social network does not solve everything. A human can troll, harass, spam, or manipulate. Verification reduces one category of nuisances, especially those relying on AI or mass automation. It does not guarantee a healthy debate. It shifts the battle to moderation and product incentives.

OpenAI’s bet, if confirmed, is also a user experience bet. Scanning one’s iris to post a message is a strong barrier. Even if the goal is noble, most people hate friction. The project will therefore have to offer an immediate and tangible benefit: less spam, more useful conversations, and a feeling of a “real world” online.

Remains a political question, in the broad sense. When the company popularizing AI also offers the system that distinguishes humans from machines, it occupies both ends of the same pipe. Some will see coherence. Others a conflict of interest or a concentration of power. This debate will not be theoretical: it will be part of the launch, if it happens. Vitalik Buterin also has other plans to save online anonymity.