Sora by OpenAI: A TikTok-Like App for AI-Generated Videos

OpenAI today officially launched its new Sora app, but for now, it’s only available by invitation to users in the United States and Canada.

The app takes on a TikTok-style interface, where videos are displayed vertically and can be swiped through. With it, users can generate AI-powered videos of themselves or friends and publish them directly within the platform. To create these videos, users need to upload their own video and audio recordings, which are then integrated into different backgrounds and scenarios using a feature called Cameo.

Of course, this opens the door to potential misuse—such as generating harmful or misleading videos. OpenAI has addressed this by requiring that users cannot impersonate someone else, and that any generated Cameo requires explicit consent through the initial video and audio upload.

The app’s launch coincides with the debut of Sora 2, OpenAI’s latest text-to-video-and-audio generation model. The updated model is said to be more advanced, producing outputs that follow the laws of physics and natural motion more closely. Currently, the Sora app is being rolled out first to ChatGPT Pro subscribers, hinting at a social platform that might gain traction much like the “Ghibli-style” image trend that spread widely earlier this year.

Sora also arrives in the wake of similar experiments from other tech giants. Elon Musk’s Grok app recently added video generation, while Meta introduced Vibes, a feature for exploring AI-generated videos inside its ecosystem. However, these launches have not been without criticism, with many raising concerns about the flood of low-value content, risks of identity misuse, and the growing problem of shrinking attention spans in the age of AI-driven entertainment.