OpenAI launched its new social video app Sora last week, bringing AI-driven video creation to the users’ doorstep. The app, restricted to iPhones in the US and Canada, operates on an invite-only system. Early invitees can also forward four extra invites to their friends. An Android launch has not been announced yet.
Unlike traditional short-video platforms such as TikTok or Instagram Reels, Sora is built around AI-generated content. Users can record a short video of themselves, called a “cameo,” which others can then use to generate fresh clips using OpenAI’s upgraded Sora 2 model. This cameo feature allows friends, or even the wider community if permitted, to add someone into their own videos. Significantly, the individual whose face is featured is still a “co-owner” of said video, and can delete or withdraw access whenever they choose.
The app has a “Remix” option to prompt users to engage with current clips and trends, but the produced videos are limited to 10 seconds for now. To ensure safety, OpenAI has placed limitations to avoid creating explicit or radical content. The company has also excluded public figures from being featured in AI-generated videos unless they upload a cameo personally and give permission for it to be used.
The app launch is accompanied by the rollout of Sora 2.0, the most advanced video generator from OpenAI to date. According to the company, the new model enhances physical realism, audio-video sync, and accommodates more sophisticated storytelling through multiple shots.
In a press conference, OpenAI staff explained the endeavor as one that has the capacity to be the “ChatGPT moment for video.” The analogy is meant to demonstrate the firm’s hopes that Sora can revolutionize the way individuals work with video production, just like ChatGPT changed typical use of AI in text.
Still, the deployment is already being met with controversy. A Wall Street Journal report found that OpenAI has notified studios and talent agencies that their copyrighted content could be included in Sora-generated videos unless specifically exempted from it. The company, for its part, maintains that it views this as a manner of fan engagement and creativity, the same way users remix images and memes.
With that being said, OpenAI has maintained a strict line on personal likeness. People cannot be created without having their cameo uploaded, so that individuals have total control over whether and how they are used in any content created with the app.
The application will later move away from the invite-only strategy. OpenAI has also indicated that a Pro version, web access, and developer API are in the works, paving the way for broader integration and business use.