OpenAI has unveiled an updated version of its Sora AI video generator, initially introduced late last year. The new app, also named Sora, is now accessible for iPhone users and aims to blend social media features with AI technology. This invite-only platform offers a video feed reminiscent of TikTok, allowing users to browse and create short clips that can be transformed into AI-generated deepfakes, contingent upon users’ consent.
During a press briefing on Monday, OpenAI employees referred to this launch as a potential “ChatGPT moment for video generation.” Currently, the Sora app is limited to users in the United States and Canada, with plans to expand to additional countries in the future. New users who gain access will also receive four invites to extend to friends. There is no announcement yet about a possible Android version.
Users of Sora can grant permission to friends, or even a broader audience, to create “cameos” that use their likeness via the platform’s new video model, referred to as Sora 2. According to OpenAI representatives, individuals whose images are utilized in these videos maintain co-ownership of the resulting content, with the option to delete or restrict access at any time. Similar to TikTok, the Sora app includes a “Remix” feature, though it currently restricts users to generating clips of no longer than 10 seconds.
Currently, OpenAI enforces strict limitations on the generation of videos featuring public figures. As stated during the briefing, such personalities cannot be included unless they have uploaded a cameo and provided consent. OpenAI clarified that this requirement applies universally; only users who upload their own cameos can have their likeness used. The company also mentioned that it is not possible to generate explicit or extreme content on the platform at this time.
