Meta unveiled an artificial intelligence to convert text to video

In recent months, artificial intelligence systems dedicated to converting text into images have been in the headlines. Meanwhile, artificial intelligence researchers working at Meta company have unveiled an intelligent system for converting text to video. This system is called Make-A-Video and by receiving a description of the scene, it can create a short video related to the written text. The artificiality of these videos is clearly evident, but in any case, the same system reports great progress in this field.

Mark Zuckerberg, the CEO of Meta, described this system as an “amazing development” in the published post and said: “Production of video by artificial intelligence is much more difficult than images because in addition to creating each pixel, the system in question must also predict the process of pixel change. Slow down.”

The clips are no longer than 5 seconds and have no sound, but they cover a wide range of requests. In the continuation of this article, you can see some of the clips published by Meta. Considering that no one outside Meta still has access to this system, probably the best clips have been selected by this company.

As we said, even though these videos are not very attractive at the moment, due to the rapid development of artificial intelligence systems related to image creation, there is no doubt that video creation systems will improve significantly in the near future. Meta wrote in the published post that such a system will have many uses for artists. But in the meantime, these types of systems draw a worrying prospect. For example, they will probably be used for advertisements and fake videos in the future.

It should be noted that Meta is not alone in the development of artificial intelligence systems related to this field. For example, several months ago, a number of Chinese researchers unveiled an artificial intelligence model for converting text to video. In the article published by Meta, the challenges related to creating videos longer than 5 seconds, videos based on diverse scenes and higher resolution are also mentioned.

Source: The Verge