AI text-to-image generators made headlines in recent months, but researchers are already moving on to the next challenge: AI text-to-video generators.
Facebook’s parent company, Meta, as well as Tsinghua University and the Beijing Academy of Artificial Intelligence (BAAI) released text-to-video models.
As the name suggests, it consists of an AI model that allows users to type in a description of a scene, from which it generates a short video matching their text.
The videos look artificial as of yet, with blurred subjects and distorted animation, but still represent a significant development in the field of AI content generation.
It’s harder to generate video than photos because besides correctly generating each pixel, the code needs to also predict how they’ll change over time.
The current output of the text-to-video models are short clips and contain no audio but span a large range of prompts.
This next generation tool could be invaluable for presentations, content creators and artists. Fast, visual and dynamic. It is a cool advancement for many aspects of communication. Just write down a sentence and create a video representing what you want to convey.