Mark Zuckerberg surprised everyone with the announcement of Make-A-Video yesterday, an artificial intelligence-powered system one-upping picture generators DALL-E and Midjourney, turning text into original video footage.
“You give it a description and it creates a video for you,” Zuckerberg wrote in a Facebook post.
The boom of AI art hasn’t gone unnoticed by Meta. From TikTok hosting their own DALL-E copycat to a piece of AI art beating real artists in a competition, the fascination with what the technology can do is growing.
Much like the text to image generators like DALL-E, Zuckerberg said the Meta team fed its new AI system descriptions like “a teddy bear painting a self-portrait”, “a spaceship landing on mars” and “a baby sloth with a knitted hat trying to figure out a laptop”.
The videos accompanying the prompts are genuinely very impressive. The AI system created high-quality, accurate and unique clips.
The CEO explained that it’s much harder to generate video than photos due to the fact the system goes beyond correctly generating each pixel.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below formBy GlobalData
“It also has to predict how they’ll change over time,“ Zuckerberg said.
Make-A-Video apparently solves this issue by adding a layer of unsupervised learning into the system.
This enables it to “understand motion in the physical world and apply it to traditional text-to-image generation.”
Zuckerberg claimed that Meta will be planning to share a full demo in the future.
Julian Vizard, a creative partner at creative agency St. Luke’s, believes Make-A-Video could be Meta’s ticket to catching up with TikTok’s meteoric rise.
“Snapchat filters already use machine learning to turn you into a Pixar-like character and TikTok recently released a filter that features DALL-E like AI image generation,” Vizard told Verdict.
“However, accessible AI video generation is a game changer and if Meta can integrate it into Reels, and gain a healthy first-mover advantage [then] it would give Meta a standout tool for creators to make unique content and make the platform look like they are ahead again.”
Vizard added that it was inevitable that Meta would join the space race of social innovation, along with Snap and Twitter.
Meta has been contacted by Verdict for comment.
GlobalData is the parent company of Verdict and its sister publications.