Instagram is making new artificial intelligence (AI) features that will allow users to reimagine their videos. AI-powered video editing tools are built on Meta’s movie Gen AI model, an early research project that can generate videos from text prompts. On Instagram, the feature does not allow users to generate videos from scratch, but instead will be videos taken with cameras and feature real humans. Users will be able to change videos, such as changing outfits, backgrounds, and even overall appearance.
The new feature was teased by Instagram head Adam Mosseri. In the long video, he showed some of the features of the new AI feature and explained how it works. It is worth noting that AI video editing tools are under development, and Mosseri emphasizes that shipments will be available next year.
In the video, the tool can change Mosseri’s outfit, add a gold chain around his neck, and even add hippos to the pool in the background. The tool also shows combining some of these effects to make the Instagram head seem to sit next to snowy mountains and in the desert landscape.
Mosseri subtitled the video, revealing that the AI video editing tool is powered by Meta’s movie AI model, which was previewed in October. It is worth noting that this is a multi-modal model with four features – video generation, personalized video generation, precise video editing and audio generation. It can also combine video and audio generation to generate video with synchronous audio. Currently, it is not available in the public domain.
“I’m very excited about Movie Gen, our early AI research model that will let you change any aspect of your video using simple text prompts. Hopefully, bringing it to Instagram next year,” Mosseri wrote in the title.
It seems that the underdeveloped AI feature borrows only the precise video editing feature of the movie general described in the announcement post, and the model described in the notification article can perform local editing, such as adding, removing or replacing elements, and making global changes, e.g. Background or style modification. The AI model also retains the original content and only targets the relevant pixels.
