Via new technology from Nvidia, AI-generated GIFs could be the next big thing to take over the internet.
Nvidia recently published a paper entitled “High-Resolution Video Synthesis with Latent Diffusion Models” based on experiments done at Nvidia’s Toronto AI Lab where they’re using Stable Diffusion to try and build out a tool that can create moving art from a text prompt. Basically, the idea is that you can type in a phrase and see that phrase turned into a moving piece of art, i.e. a GIF.
According to Nvidia, the company is able to generate moving images that are roughly 4.7 seconds long at a 1280 x 2048 resolution; though, the tool is also reportedly capable of creating longer videos at a 512 x 1024 resolution, too. While the tool is capable of generating this moving art based on a text prompt, reportedly there is still more work to be done to eliminate distracting artifacts.
What would such a tool be used for? Well, the biggest usecase could well be the creation of GIFs. AI-generated images and selections from AI chatbots have already taken over the internet, and that’s to say nothing of deepfake videos, either. AI-generated videos could blend the best of both worlds and offer up a moving piece of art that’s easy to distribute and share around without having to commit to watching a video.
Plus, GIFs are huge when it comes to memes online, so being able to create these on demand with the help of AI it’s easy to see spreading like wildfire on a platform like Twitter. Though, as of now, this is very much all still speculation. It’s unclear how Nvidia will use this tool, and it’s unclear how easy it will be to use by the average consumer.
