Meta introduced its new platform, Movie Gen, which has the capacity to produce high-resolution videos with artificial intelligence. Movie Gen allows users to create videos with just text input. However, this technology has not yet been put into general use. This announcement, which comes months after the introduction of Sora, OpenAI’s text-to-video creation tool, increases competition in the artificial intelligence field.
Movie Gen allows text-based video production as well as editing of existing videos and images. For example, it has the ability to turn a still photo into a moving scene or replace the content of an existing video with different visual and sound effects. This technology allows videos to be created in various ratios and formats. In addition, the added sounds are produced by artificial intelligence, and ambient sounds, effects and music are added in sync with the images.
Meta gave examples of Movie Gen’s capabilities, showing how it transformed a headshot into a moving image of a woman drinking in a pumpkin field. During this process, the AI demonstrated its capacity to create a realistic video from a still image. Thus, new stories can be created by turning fixed images into dynamic videos.
Movie Gen’s editing ability
One of the most notable features that Movie Gen offers is the ability to edit existing videos with the help of artificial intelligence. Users can change the costumes of characters in a video, replace backgrounds with different locations, or add additional visual elements. An example shared by Meta shows how a running character is transformed by different scenes. In one frame, the runner runs with pom-poms in his hand, while in another frame there is a desert landscape in the background. In another scene, the character wears a dinosaur costume. All these changes can be made using text commands only.
In the world of technology, video production with artificial intelligence has become a rapidly developing field in recent years. While the use of artificial intelligence in visual production began to become widespread two years ago, video production technologies have also gained great momentum in the last six months. Google, OpenAI and smaller initiatives continue to make important steps in this field. OpenAI’s video creation model called Sora is not yet available, but the transfer of an executive working on Sora to Google this week shows that the competition will increase even more.
Meta’s director of product development, Chris Cox, said in a statement on the Threads platform that Movie Gen is not yet ready to be released as a commercial product. Cox stated that this technology is high cost and video creation time still takes a very long time. So, it may be a while before Movie Gen becomes widely available.
Some ethical debates regarding artificial intelligence-supported video production continue to come to the fore. For example, an artificial intelligence startup called Runaway allegedly trained its video production model on thousands of YouTube videos. It is stated that this situation violates YouTube’s terms of use and may cause copyright issues. Although Meta stated that “licensed and publicly available datasets” were used to train Movie Gen, it did not share details about which sources were used.
On the other hand, creative professionals in the cinema and media world express concerns about how artificial intelligence technology will affect their profession. Filmmakers, photographers, writers and actors think that content produced by artificial intelligence threatens their professions. These concerns also played a central role in last year’s Screen Actors Guild of America (SAG-AFTRA) and Writers Guild of America (WGA) strikes.