Meta, Facebook’s parent company, teased a new artificial intelligence tool on Friday that allows users to make 16 second videos based on text prompts. The new tool, called Movie Gen, pits Meta against companies like OpenAI in a battle of text-to-video AI features.
Movie Gen will allow users to create clips based on what they enter its text prompt. If you typed “A fluffy grey and white koala bear surfs on a yellow board,” as Meta shared as an example on its blog, here’s a screenshot of what you’d get:
Users will also be able to create images and audio using Movie Gen’s text prompts.
If you’re dying to use Mark Zuckerberg’s new AI feature, though, you’ll have to wait a bit. Movie Gen is only available to some company employees and “outside partners, including some filmmakers,” according to Bloomberg. Meta is planning on a wide rollout of Movie Gen in 2025, with the feature being available on its handful of apps, including Facebook and Instagram.
The company gave itself a pat on the back for Movie Gen’s realistic images and videos, saying it “outperforms similar models in the industry across these tasks when evaluated by humans.”
It’s an increasingly crowded industry too. OpenAI’s Sora text-to-video tool can make clips that are up to 60 seconds long, and Google debuted Veo, its own text-to-video tool earlier this year. A number of startups are on this corner, too, like Stability AI, a London-based company that just added James Cameron to its board of directors last month.
The filmmaker heralded text-to-video tools as the “next wave” of creation that’ll “unlock new ways for artists to tell stories in ways we could have never imagined.”