Meta has announced a new AI tool for video creation called Movie Gen, powered by artificial intelligence (AI). This tool can create high-definition videos with sound. Although Movie Gen was revealed today, it is not yet available for the public to use. This announcement comes just months after OpenAI introduced its own text-to-video model called Sora.
What Movie Gen Can Do
Movie Gen allows users to type in text, which the AI uses to automatically create new videos. It can also edit existing videos or still images. According to The New York Times, the audio in these videos is generated by AI as well, providing sounds that match the visuals, like background music and sound effects. Users can generate videos in different aspect ratios, depending on their needs.
In addition to making brand-new clips, Movie Gen can create custom videos from images. It can even take a pre-existing video and modify various aspects of it. For example, Meta shared a demonstration where a still image of a woman was transformed into a video of her sitting in a pumpkin patch while sipping a drink.
Editing Capabilities
Movie Gen also offers editing features for existing footage. It can change the style, transitions, and even add new elements that weren’t in the original video. One example Meta provided showed an animated runner. In one version, the runner is holding pompoms; in another, the background is changed to a desert; and in a third version, the runner is dressed in a dinosaur costume. These changes are made using simple text prompts.
The Competition and Concerns
Since powerful AI image and video generators have become popular in the last two years, many tech companies are working on similar tools. In just the last six months, major companies like Google and OpenAI have been developing their own versions. OpenAI’s Sora was first announced in February but is still not publicly available. Recently, a key developer from Sora left to join Google.
Chris Cox, Meta’s chief product officer, mentioned on Threads that they are not ready to release Movie Gen as a product just yet because it is still costly and the time it takes to generate videos is too long.
As AI technology advances, it raises concerns about ownership and how it might be misused. Reports indicate that some AI startups, like Runaway, trained their video generators on thousands of videos taken from YouTube, which YouTube’s CEO Neal Mohan has stated violates the platform’s rules. Meta claims it trained Movie Gen using a mix of licensed and publicly available data but did not specify exactly what sources were used.
Creatives such as filmmakers, photographers, artists, writers, and actors are worried about how these AI generators might impact their jobs. This issue has even been a key topic in several labor strikes, including the historic Hollywood strikes involving the Screen Actors Guild – American Federation of Television and Radio Artists (SAG-AFTRA) and the Writers Guild of America (WGA) last year.
