Meta has announced an AI-powered video generator, dubbed "Movie Gen", capable of producing high-definition footage complete with sound, according to a recent blog post by the company. However, public access to Meta's Movie Gen is still on hold.

The announcement from Meta comes several months after OpenAI announced its own AI-based video generator, Sora.

According to Meta, the tool can create entirely new videos from text prompts, as well as edit existing footage or still images. The AI can also generate audio, matching the video with ambient sound, effects, and background music. Videos can be outputted in various aspect ratios, providing flexibility for different platforms.

In one demonstration, Meta showcased how Movie Gen could transform a simple headshot of a woman into a video where she appears to sit in a pumpkin patch, sipping a drink. 

Image: Meta.

The technology also allows for more complex edits, such as altering the style of an existing video or adding new elements. Meta demonstrated this with a clip of an illustrated runner, who was transformed in different scenarios such as waving pom poms, running through a cactus desert, or even dressed in a dinosaur costume by using text commands.

Image: Meta.

Read more

Watch the first-ever commercial ad made by OpenAI's Sora