In a development that should surprise no one, OpenAI is preparing to release a standalone social app powered by its upcoming Sora 2 video model, Wired reports. The app reportedly “closely resembles” TikTok, with a vertical video feed and swipe-to-scroll navigation. The catch? It will only feature AI-generated content; there’s apparently no option for the user to upload photos or videos from their phone’s camera roll.
Wired adds OpenAI will limit Sora 2 to generating clips that are 10 seconds long or shorter for use inside of the app. It’s unclear what the model’s limit will be outside of the app. TikTok, following an original limit of 15 seconds, allows people to upload clips that are up to 10 minutes long. The new app is also said to include an identity verification tool. Should a user take advantage of the feature, Sora 2 will be able to use their likeness in videos it generates. In turn, that means other people will be able to tag those users and use their likeness when they go to remix one of their videos. As a safety precaution, OpenAI will push a notification to users whenever their likeness is used by someone else, even in situations where someone makes a video but never posts it to the app’s feed.
According to Wired, the software will refuse to generate some videos due to copyright restrictions. However, it’s not clear just how robust these protections will be, with The Wall Street Journal reporting OpenAI will require rights holders to opt out of their content appearing in videos Sora 2 generates.
As for why OpenAI would release a social media app, Wired suggests the company saw an opportunity after President Trump repeatedly extended the deadline for ByteDance to bring TikTok’s US business under American control. By adding a social component to Sora, OpenAI may also be hoping to dissuade people from trying other models since leaving its new app would mean abandoning whatever community forms around it.
