Meta has unveiled its new AI technology “Movie Gen,” capable of generating videos and audio from text descriptions. The system comprises several models that together can produce high-resolution videos up to 16 seconds long, along with matching audio effects. According to Meta, Movie Gen surpasses comparable offerings from competitors like OpenAI and Google in terms of quality and consistency.
The technology allows users to create personalized videos by uploading a photo of themselves. Additionally, existing videos can be precisely edited, such as adding or changing objects through text instructions. For audio generation, Meta promises realistic sound effects and mood-appropriate background music.
Meta plans to introduce Movie Gen on Instagram in 2025. The company emphasizes that the technology is intended to support, not replace, creatives and filmmakers. However, critical questions about the data foundation have been raised: Movie Gen was trained on millions of online videos and images, though it remains unclear whether this data was licensed or freely available.
Sources: Wired, Bloomberg, VentureBeat