Making and completing your own videos is about to get a boost of the unreal, as Adobe readies its own AI video solution for content creators.
Every week, there’s something new about AI. The year’s most obvious buzzword and trend sees a new story more or less every week, and much of it is around saving you time. As to whether it truly does save time is another question, but the idea is there.
Save time reading by letting your phone summarise your emails. Save time making images by throwing a text prompt into an AI system. Pretty soon, you might be able to save time by letting an AI make videos for you to assemble, extend, and complete projects, so to speak.
That’s just part of what Adobe has announced this week, as Adobe joins ChatGPT’s OpenAI and Meta’s efforts to deliver AI-based video from mere prompts.
Adobe’s is trained on what the company calls media “safe for commercial use”, which means you shouldn’t be able to pop out a 1:1 clone of The Terminator or any other big name movie, simply because the AI hasn’t been trained on something it didn’t have permission for.
That’s a huge problem in artificial intelligence at the moment, and affects all AI concepts. You can’t just use someone else’s intellectual property to train your AI model, and yet that’s happening, such as how you can make AI music from nothing sound like big name artists.
You won’t be able to necessarily do that with Adobe’s video platform, and it will use its model to create videos and sounds you can use based on your text or image prompts, thanks to its Firefly AI system on its website.
However, Adobe will go beyond this with an integration in Premiere Pro beta, allowing videos to be extended using Firefly Generative Extend, essentially making up frames out of nothing in case your video files are too small.
“We’re giving the creative community a powerful new brush to paint the world by putting unprecedented power, precision and creative control in their hands,” said David Wadhwani, President of Digital Media at Adobe.
Adobe’s additions go beyond AI in video, with Photoshop set to deliver a way to remove distractions in photos and art quickly and easily, a beta of something called “Generative Workspace” to create and churn out colour variations for art you may be working on, and a faster way to convert images to vector using “Enhanced Image Trace” inside Illustrator.
The additions are gradually rolling out, but will be available for some first via a waitlist, with the initial pricing on AI video generations free while in public beta. Adobe notes that eventually, AI video will cost money, with pricing available when the beta period is over.
