Google’s New Video Tool
If you’re interested in making videos simply by typing a text prompt, Google may soon have just the tool for you. It’s called Google Lumiere.
Lumiere uses artificial intelligence (AI) to generate videos based on a written description. A promotional video shows a text prompt reading, “Astronaut on the planet Mars making a detour around his base.” The resulting video shows a person in a space suit walking around the Red Planet. A prompt asking for a dog in sunglasses driving a car leads to a video showing a dog in sunglasses driving a car. These aren’t existing videos that Google found. They’re generated by AI, so the possibilities are endless.
Google isn’t the first company to develop this sort of technology. Many other companies have text-to-image creators, and a few have developed text-to-video generators like Lumiere. But many people who have watched Lumiere video clips report that, while Google’s product isn’t perfect, it creates the most realistic AI videos yet.
Omer Bar-Tal (Google Research, Weizmann Institute), Hila Chefer (Google Research, Tel-Aviv University), Omer Tov (Google Research), et al. “Lumiere: A Space-Time Diffusion Model for Video Generation.” 2024.
Three examples of Google’s text-to-video technology.
Lumiere has many additional capabilities. Users can turn images into videos, or just animate one part of an image. They can also edit existing videos and even repair damaged videos or fill in parts of a video that are missing.
As of February 2024, Lumiere was not yet available to use. But based on the video clips that Google has released, it’s easy to imagine what could be created in the future.
