OpenAI, the AI company that created ChatGPT, is also supposedly working on a new social network platform that will only permit accredited human users and significantly limit the activity of automated bots, according to individuals involved in the project.
Assuming it becomes a reality, the network has the potential to transform the way social platforms manage identity verification and online authenticity.
The project is still at the early phase and is under development by a small internal team of less than ten individuals. A timeline to a public launch is not confirmed, and OpenAI has not officially responded to the project.
The essence of the social network is to make sure that all accounts are owned by real people, not by automated systems or intelligent AI-controlled accounts that imitate human behaviour.
This is amid a time when most mainstream apps, such as X (previously Twitter), Instagram, and Tik Tok, have long-standing problems with fake accounts, spam, and AI-bot chatter flooding user interactions.
The difference with OpenAI is that strong identity checks, such as biometric verification, may be used to ensure that the user is a human being. Instead of using traditional identifiers like phone numbers or email addresses, the platform may ask users to verify who they are using technology like Apple Face ID.
Even the World orb, a human iris scanning device created by Tools For Humanity, an organization that was chaired by OpenAI CEO Sam Altman.
Advocates assume that biometric authentication will make it much more difficult to have bot networks gain a foothold on the site. Nevertheless, the concept already raised eyebrows among privacy activists, who caution that biometric information is particularly dangerous when it is leaked or otherwise misused.
In contrast to passwords or phone numbers, biometric identifiers are impossible to change once compromised, and the question of data security and the privacy of users over time emerges.
Although much of the description of the platform and functionality is hypothetical, according to insiders, a user would still be able to engage with AI technology to create content, like photos and videos, in the same trend as on Instagram and TikTok.
That would make the new network a direct competitor to the existing social platforms that are already starting to incorporate generative AI into services.
There is a partial motivation to create a bot-free environment based on general discontent at the prevalence of automated accounts controlling online discussions. This has been especially apparent on X, where increased bot operations coincided with major layoffs in the trust and safety divisions of the platform owned by Elon Musk.
Although attempts are made to eliminate fake accounts, such as the removal of approximately 1.7 million bots in 2025, users still complain about spam and artificial interaction.
The management of OpenAI has already recognized these issues publicly. The theory of the dead internet posits that an increasing part of online interactions occurs due to AI and not real people; Sam Altman has made comments about this theory. His words indicate a wider discomfort with the loss of real conversation in online forums.
The history of OpenAI in consumer applications supports the company. ChatGPT racked up 100 million users in two months after launch and now serves hundreds of millions of users around the world, and its AI video app, Sora, has over one million downloads in less than five days.
However, creating a new social network is a challenging first step, particularly in a saturated market where giants like Instagram, Tik Tok, Threads by Meta, and up-and-comers like Bluesky operate.
Also Read: TikTok Influencer Alina Amir’s Viral MMS Deepfake Explained as She Breaks Her Silence
To date, OpenAI has kept the project publicly silent, and it is not certain that it ever will be released. However, assuming the company does proceed, it will be trying a daring theory: that users actually desire a digital realm in which they can have certainty that they are communicating with actual people, not with lines of code masquerading as profiles.
