Key Points
- OpenAI has released Sora 2, a major upgrade to its AI video and audio generator.
- The new model is paired with a new social app, also called Sora, for creating and sharing videos.
- A “cameos” feature lets you create a digital likeness of yourself to use in the videos.
- Sora 2 is more realistic than its predecessor because it better adheres to the laws of physics.
OpenAI has just launched Sora 2, a major upgrade to its AI video and audio generator, accompanied by a brand-new, TikTok-style social app. The new app, also known as Sora, allows users to create and share AI-generated videos featuring themselves and their friends.
Sora 2 boasts a major improvement over its predecessor: it’s much better at following the laws of physics, making its videos far more realistic. OpenAI says that in older models, a basketball that missed a shot might just “teleport” into the hoop. With Sora 2, the ball will bounce off the backboard just like it would in real life.
The centerpiece of the new Sora app is a feature called “cameos.” This allows you to upload a short video of yourself to create a digital likeness, which you can then insert into any AI-generated scene. You can also share your “cameo” with friends, permitting them to include you in their own creations.
Videos created in the app can be shared on a social feed, which resembles TikTok or Instagram Reels. While the app is currently invite-only, ChatGPT Pro users will be able to try out the new Sora 2 model without an invite.
The launch of a social platform built around AI-generated likenesses raises significant concerns about safety. While OpenAI states that users can revoke access to their “cameo” at any time, the potential for misuse and the creation of non-consensual videos poses a significant risk that the company must navigate carefully.
