OpenAI is secretly building a social network, master plan to catch bots leaked

OpenAI may be quietly preparing its boldest consumer-facing experiment yet, and it has little to do with chatbots answering questions. People familiar with the matter told Forbes that the company is working on an early version of a social network designed around a single promise, which is basically conversations led by real people and not automated accounts pretending to be human.
The idea comes at a time when many popular platforms feel increasingly crowded with fake profiles, spam replies and AI-generated chatter. Insiders say OpenAI wants to tackle that problem at the root by building a space where every account belongs to a verified person. The project is still in its infancy and is reportedly being built by a very small team of fewer than ten people, with no public launch plan in sight yet.
What makes OpenAI’s approach different is how seriously it is considering identity checks. Unlike existing social apps that rely on phone numbers, email IDs or behavioural patterns, this network could ask users to prove they are human through biometric verification. Options discussed internally include Apple’s Face ID and the World Orb, a device that scans a person’s iris to create a unique digital identity. The World project is run by Tools for Humanity, a company founded and chaired by OpenAI CEO Sam Altman.
Such a system would make it extremely difficult for bot networks to exist on the platform. At the same time, it has already raised alarms among privacy advocates, who warn that biometric data cannot be changed if misused and could create serious risks if leaked or mishandled. Sources say these concerns are part of ongoing internal discussions.
While details about the app’s features remain limited, people aware of the project say users would likely be able to use AI tools to create content such as images or videos. That would place OpenAI in direct competition with platforms like Instagram and TikTok, both of which are already leaning heavily into AI-powered creation tools. Instagram alone has over three billion monthly active users, underlining just how crowded the space already is.
The motivation behind the project appears closely tied to the current state of social media, particularly X. Bot activity has long been an issue across platforms, but it became far more visible after Elon Musk took over Twitter, renamed it X, and drastically reduced its trust and safety workforce. Although the company removed around 1.7 million bot accounts in 2025, spam and automated replies continue to frustrate users.
Altman has spoken openly about this problem. In posts on X, he has said that discussions around AI on social platforms now feel strangely artificial, filled with accounts that do not seem human. He has even pointed to the so-called dead internet theory, noting that large language model–run accounts appear to be far more common than they were just a couple of years ago.