Saturday, October 01, 2022
The results of the tool are quite impressive and in the words of the CEO "pretty amazing".
Called Make-A-Video, artificial intelligence in the system uses existing images to "understand motion in the physical world and apply it to traditional text-to-image generation."
"You give it a text description and it creates a video for you," Zuckerberg said in a statement.
Previously, it has been very difficult to generate video because not only does the AI have to generate pixels but also predict how they would change with time.
This problem is solved by adding the feature of "unsupervised learning" to the system.
The tech giant is yet to reveal when the tool will be available for the masses or if there will be restrictions like subscribing or signing up to the system. The company did say that it would share demos soon.
With the possibility of AI being misused for disinformation and deepfakes, there are always ethical dilemmas around the subject.
To solve this, Meta said that the videos generated from Make-A-Video will have a watermark.
"For now we will continue to analyze, test, and trial Make-A-Video to ensure that each step of release is safe and intentional," Meta said.