Meta Unveils ‘Make-A-Video’ AI Text-To-Video Generator

Meta has introduced a new artificial intelligence (AI) system called Make-A-Video. This AI system allows users to create short video clips by entering a text description of the desired scene. The featured announcement is seen as an advance in generative technology research aimed at giving creators more creative control over artificially intelligent image generation. With the help of this new invention, Meta has taken the technology a step further by incorporating text-to-video generation capabilities alongside text-to-image capability. However, the company has not yet disclosed access to the users of the model.

The videos generated by the prompt are five seconds or shorter and do not contain audio. However, Meta has made a promise regarding the variety of prompts supported by the model.

Via the blog post, the company mentioned that as part of its commitment to “open science,” it will share details about the research behind the latest generative artificial intelligence technology. The company has various plans on the list; The main plan is to release a demo experience for users.

Generative AI research encourages creative expression by giving people the tools to create new content quickly and easily.” “With just a few words or a few texts, the Make-A-Video tool can turn imagination into implement reality and create unique videos that can be full of vivid colors and landscapes,” the Meta mention reads.

In the research paper describing the model at work, the company noted that the Make-A-Video demo model uses image pairs, captions, and uncaptioned video footage extracted from WebVid-10M and HD VILA-100M records originates. This includes stock video footage created by sites like Shutterstock and scraped from the internet, totaling hundreds of thousands of hours of footage.

Meta CEO Mark Zuckerberg described the work, using Facebook as an example, as “amazing progress.” He also mentioned that it’s much more difficult to create videos than images because, in addition to collecting every single pixel, the system also has to determine how they will change over time.

However, worrying issues have been raised around generative AI media. There are some of the suggestions received that could lead to an increase in misinformation, propaganda, and non-consensual pornography. The same has been observed in the case of AI image-generating systems and deepfakes. The company wants to think better about how to create such generative models, so it plans to restrict access to them. However, a timeline for the demo experience and clarity on how access would be restricted is not yet known.