Meta has launched Meta.ai, an AI-specific site that has a cool hook that its competitors don’t offer: It can generate images in real time, and even animate them on demand.
There is a catch, however: Meta would really like to continue improving Meta.ai, and to do so it’s only offering image generation if you sign into your Facebook account.
Meta joins other LLMS or AI chatbots like Google Gemini, Microsoft’s various flavors of Copilot, Anthropic’s Claude AI (used within Discord), and other sites offering AI solutions. Meta.ai feels like more of the same, though with some limitations: It can’t accept uploaded documents, but it can summarize websites or web pages. Of course, it has creative purposes, too: It can also be used to write or rewrite text, as many other services can as well.
Technically, Meta is upgrading Meta AI as well as launching the site itself. Meta chief executive Mark Zuckerberg announced the addition of the Meta.ai site on Threads — which is also owned by Meta — as well as the upgrade to the LLaMA 3 model. Meta launched LLaMA in February 2023, and publicly released the 65-billion parameter model. LLaMA 3 will be open-source, as well, Zuckerberg said.
But there’s more: “We’re making Meta AI easier to use by integrating it into the search boxes at the top of WhatsApp, Instagram, Facebook, and Messenger,” Zuckerberg added. In other words, expect Meta AI to be all over your social media.
The coolest thing about Meta.ai is its image generation — which is also its most problematic, if you don’t really want to tell Facebook what you’re making images of, or don’t subscribe to Facebook or Instagram in general. But the image progression might be unlike anything you’ve ever seen. If you’re familiar with how Google or Bing starts auto-generating suggestions as you type, Meta does almost the same thing — except it’s drawing the actual pictures as you type, too.
Mark Hachman / IDG
It does so at blazing speed, even if the eventual image doesn’t necessarily come out as sharply as some of the final images. When completed, Meta generates four 1280×1280 images, which can be highlighted or shown off at a larger size. They’re all downloadable, and all come watermarked with an “Imagined by AI” watermark. The images can also be edited after completion, which makes changes to the finished product.
The other feature Meta offers, however, is the ability to animate each of the resulting images This isn’t anything like Runway or Sora; instead, Meta.ai generates a few frames of video before and after the original image, like a GIF. The process took at least 30 seconds to a minute on my screen. (On a second try, it seemed to lock up and didn’t complete after several minutes.)
Meta.ai
One has to wonder if Meta, as the owner of Facebook, will allow AI art to be generated on Facebook as a way of generating misinformation. A quick effort to try and draw Donald Trump generated a non-specific error message. But any image that you create can still be downloaded and the watermark cropped out.
Meta.ai has enough to offer that we’d recommend that you try it out. But I think it’s fair to say that its long-term impact won’t be on Meta.ai, but on the other properties Facebook owns.
