OpenAI is reportedly preparing to launch a social media app alongside its upcoming Sora 2 video model, according to Wired. The app is said to mimic TikTok’s vertical video feed and swipe-to-scroll interface, but with one major twist: all clips will be AI-generated. Users won’t be able to upload their own videos or photos from a phone’s camera roll.
Sora 2 will initially be limited to creating clips no longer than 10 seconds within the app, though it’s unclear if the model will have longer capabilities outside of it. By contrast, TikTok has steadily expanded from its original 15-second limit to allow videos of up to 10 minutes. OpenAI’s version appears to be focused on fast, snackable content that leans entirely on generative AI rather than user-shot footage.
The app will also feature an identity verification tool. Users who opt in can allow Sora 2 to generate videos using their likeness, which other people can then remix and tag. To address safety concerns, OpenAI will notify users whenever their likeness is used, even if the resulting video is never publicly posted.
Copyright protections are expected to be built in, though reports suggest the approach may be opt-out rather than opt-in, meaning rights holders will need to actively request that their material not be used in generated clips. How effective this system will be remains unclear.
The timing of the launch may be strategic. With TikTok’s U.S. operations still in limbo following repeated delays on a forced divestment deadline, OpenAI may see an opening to position its own video platform as an alternative. Tying a community-driven social layer to Sora 2 also gives the company a way to keep users inside its ecosystem, discouraging them from experimenting with competing generative video models.
Whether audiences will embrace a feed made entirely of AI-generated clips is an open question, but the move suggests OpenAI wants to push Sora beyond being just a tool and into being the foundation of a social platform.
