/*-- New Custom Menu Style --*/

Meta introduces new AI system that allows users to convert text instructions into video clips

Jonas Patrikalos

24/11/2022

-

Reading time: 2 min

The most important facts in brief
  • New video AI from Meta presented
  • It can generate videos from texts
  • The artificial intelligence should also be able to generate videos from static images
Rock your online store content!

Better rankings & more sales through perfect content for your target group

Download now for free

New Make-A-Video AI

Make-A-Video is a new AI art generator from Meta that can create short, high-quality video clips from text snippets.
The Meta Group writes in a press release: "The system learns what the world looks like from paired text-image data and how the world moves from video material without associated text. In this way, it learns how to create videos on its own, based on the data available to it. As part of our commitment to transparent research, we are sharing the details in a research paper and plan to release a demo."

Make-A-Video can produce significantly higher quality videos than was previously the case with other A.I.'s. All users have to do is formulate a rough description of the desired video. The A.I. then creates an approx. 5-second video without sound. So far, the videos generated are reminiscent of the GIFs we are familiar with, but it is assumed that the A.I. will deliver much higher quality videos in the future.

A dog wearing a Superhero outfit with red cape flying through the sky
Source: Meta. - https://makeavideo.studio/

The main features of Make-A-Video are:

  1. Generate videos from texts
  2. Generate videos from static images
  3. Generate further variants from existing videos

With Make-A-Video, Meta seems to be taking AI art into a new moving image era. If you are more interested in the topic, you can find a paper(PDF) on Meta's A.I. model online.

New Make-A-Video AI

Make-A-Video is a new AI art generator from Meta that can create short, high-quality video clips from text snippets.
The Meta Group writes in a press release: "The system learns what the world looks like from paired text-image data and how the world moves from video material without associated text. In this way, it learns how to create videos on its own, based on the data available to it. As part of our commitment to transparent research, we are sharing the details in a research paper and plan to release a demo."

Make-A-Video can produce significantly higher quality videos than was previously the case with other A.I.'s. All users have to do is formulate a rough description of the desired video. The A.I. then creates an approx. 5-second video without sound. So far, the videos generated are reminiscent of the GIFs we are familiar with, but it is assumed that the A.I. will deliver much higher quality videos in the future.

A dog wearing a Superhero outfit with red cape flying through the sky
Source: Meta. - https://makeavideo.studio/

The main features of Make-A-Video are:

  1. Generate videos from texts
  2. Generate videos from static images
  3. Generate further variants from existing videos

With Make-A-Video, Meta seems to be taking AI art into a new moving image era. If you are more interested in the topic, you can find a paper(PDF) on Meta's A.I. model online.

About the author
Jonas Patrikalos

As a shrewd media buyer, Jonas specializes in making campaigns on Meta effective for clients.

Rock your online store content!

Better rankings & more sales through perfect content for your target group

Download now for free
Jonas Patrikalos

As a shrewd media buyer, Jonas specializes in making campaigns on Meta effective for clients.

BEST CASES

Performance Marketing Insights