Syed Saif Ul Hassan 
Oct 01, 2025
Keeping up with the immense progress we've been witnessing in audio and video artificial intelligence (AI), OpenAI is replacing its Sora with an advanced audio and video generator called Sora 2.
The launch of OpenAI's Sora 2 will be accompanied by a new social app called Sora, which enables users to create and share videos of themselves and friends on a TikTok-style feed.
While Sora 2 is currently an invite-only offering and far from a stable release, OpenAI shared some tempting demonstrations of its capabilities, giving a sneak peek into its surprising realism and adherence to the laws of physics.
For example, while earlier AI models were prone to show a basketball teleporting to the hoop after a missed shot, Sora 2 realistically depicts the ball rebounding off the backboard.
The headlining feature of the Sora app is called “cameos,” allowing users to insert themselves into Sora-generated scenes. To use this feature, users need to upload a one-time video and audio recording for identity verification.
As per TechCrunch, the Sora is exclusively available in the US and Canada as of now, and is set to expand globally soon. In it, users can share their videos in a feed similar to TikTok or Instagram Reels.
It was also highlighted that the AI titan will curate content recommendations based on user activity, location, and engagement, with parental controls also being on the cards.
At launch, the Sora app will be free to encourage exploration, with plans to charge for generating extra videos during high demand.
Nevertheless, a probable challenge that is common to AI models is user safety and consent issues, as the misuse of AI-generated likenesses poses significant challenges.
