Meta has unveiled Movie Gen, a cutting-edge AI model that generates highly customizable, longer videos based on prompts. The announcement highlights Meta’s focus on advancing video-creation technologies amid growing competition from companies like Adobe and OpenAI. According to Ahmad Al-Dahle, Meta's vice president of generative AI, Movie Gen can create 16-second video snippets from a single prompt and perform precise edits, such as replacing objects within scenes.
Although impressive, the tool isn't ready for public use yet. Meta is still refining it and gathering feedback from filmmakers. The introduction of generative video tools has stirred excitement among tech giants and startups. Still, the complexity of scaling these services and ethical concerns remain a hurdle. Similar tools from competitors like OpenAI's Sora and Adobe’s Firefly Video Engine are also in development but have yet to be widely released.
Meta has been quiet on the specific data sources used to train Movie Gen, noting only that they rely on a combination of licensed and publicly available content.