Make-A-Video

Make-A-Video is an AI system developed by Meta AI that generates unique, high-quality videos informed by text prompts. The technology uses a text-to-image generation process to create its videos. Make-A-Video utilizes pictures with descriptors to comprehend the appearance of the world and how it is typically described.

Additionally, it uses unlabeled videos to learn world motion. Users have the option to generate videos that vary from surreal to realistic. Make-A-Video is a progression from static image generation to the generation of videos filled with motion from a single image or a pair.

There’s an option to generate multiple versions of the same video, increasing creative possibilities. Currently, the tool is still a work in progress, with caution applied to ensure that any release is deliberate and secure, to minimize the generation of misleading, harmful or biased content. To clearly identify its AI-generated content, Make-A-Video adds a watermark to all of its videos.