Meta has revealed a pair of new generative AI tools that will let Instagram and Facebook users make videos from a simple text prompt — or make changes to an image with an equally simple text prompt.
Known as Emu Video and Emu Edit, the new features use the same technology that powers sticker creation in WhatsApp and other image-generation capabilities inside Meta products.
Crafting a video with Emu will be as simple as telling it what you want the video to look like — this could include turning a static image into an animated GIF or short clip for Reels.
How does it work?
Mark Zuckerberg announced the products in a Reels video posted on Facebook — the short clip featured the Meta CEO playing with his dog on a patch of grass slowly turning into a panda, then a panda in the jungle, and finally a pair of dancing pandas in a disco.
Emu is a massive foundation AI model that is being integrated across Meta's social and communications products and services. The latest features add video for the first time — giving users a simple way to craft content without having to open a camera or learn how to animate.
The other exciting development is editing existing images. Emu Edit can be used to add or remove text from a photo you already have in your camera roll, it could allow you to add a new object into the image or change it completely, maybe placing your phone on the Moon.
Why do we need it?
Image generation with AI can be something of a dark art. You enter a prompt and hope it comes out the way you expect but if not you enter further prompts until it is close enough. Emu Edit is more free-form, according to Meta. It works through multiple tasks and instructions.
It can remove or add a background, change colors and the geometry of the whole image or any individual part, it can detect parts of an image and change them. Basically, it has the capacity to give even the least technical the ability to make complex changes to images.
Meta wrote in a blog post: “While certainly no replacement for professional artists and animators, Emu Video, Emu Edit, and new technologies like them could help people express themselves in new ways."
Adding that this could include an "art director working through a new concept or a creator livening up their latest reel" through to "a best friend sharing a unique birthday greeting."
We are starting to see the fruits of years of research into artificial intelligence come to the public sphere. Whether it is Google adding the ability to clone the voices of your favorite musical artists in YouTube, or Meta letting you turn your dog into a dancing panda — this is just the start.