Although AI text-to-image generators have been receiving a lot of attention lately, researchers have shifted their focus to the next challenge: AI text-to-video generators. Recently, Meta has revealed "Make-A-Video," an AI-powered video generator that can produce original video content from text or image inputs, similar to popular image-generating tools like DALL-E and Stable Diffusion.
On their blog post, Meta's Make-A-Video displays sample videos created from text prompts, such as: "a musk ox grazing on beautiful wildflowers." and "a blue unicorn flying over a mystical land." It also demonstrates Make-A-Video's capacity to animate still images. For instance, a static image of an eagle can appear to be flying when processed through the AI model. The videos you see below were created by Make-A-Video using captioned prompts to generate the video.