Google has announced a video generator based on artificial intelligence. The tool is called Lumiere and is capable of creating short videos of up to 5 seconds, writes Ars Technica.

What makes Lumiere special is that it uses the unique U-Net spatio-temporal architecture and generates the entire video duration at once.

U-Net

In addition, Google’s AI tool creates videos from text and images, generates videos in certain styles, animates certain parts of the image, and changes the colors of elements in the video.

Lumiere is currently a research project. At the same time, Google did not specify what data was used to train the model.

“We train our T2V [text to video] model on a dataset containing 30M videos along with their text caption,” the company added.

They also noted that the main goal of this work is to enable novice users to create visual content in a creative and flexible way.

However, the company recognizes the risk of technology being misused to create fake or malicious content. Therefore, they believe it is important to develop and apply tools to detect misuse.