Apple has announced an innovative new artificial intelligence tool which could enable anyone to create animations. Apple Keyframer is an AI animation generator that needs nothing other than a still image and a text instruction.
It comes just a few days after the company revealed MGIE, a way of editing photos using text commands …
Apple’s quiet work on AI
Apple has long been criticized for apparently being left behind in the AI field. People complained that Siri hasn’t evolved to keep pace with smarter intelligent assistants from Google and Amazon – a criticism which we’ve argued in the past has never been entirely fair. We’ve also argued that Apple has actually been using AI in other ways for more than a decade.
But it’s now becoming increasingly clear that the company has been quietly working on a wide range of AI projects, and simply doing its usual thing of saying nothing about them until it is ready to do so.
We saw that last week, when a research paper and demo described a way to edit photos simply by telling the AI system what we want to achieve.
Just a few days later, VentureBeat spotted that Apple researchers have published another paper for manipulating graphics using text commands. Apple Keyframe is an AI animation tool which lets users take a single image, and then use natural language text to tell the system how to animate it.
You can, for example, take a static image of the planet Saturn, and issue the command “Make the planet spin.”
You can also use it to help generate ideas. For example, a graphic containing stars and the command “Create three designs where the stars twinkle.”
The research paper sees its initial target market as graphics professionals, who would use it to quickly test out ideas before working on them conventionally.
Informed by interviews with professional animation designers and engineers, Keyframer supports exploration and refinement of animations through the combination of prompting and direct editing of generated output. The system also enables users to request design variants, supporting comparison and ideation.
Apple says Keyframer is more powerful than existing tools which aim to do the same thing, because it’s possible to refine ideas – either using more text prompts, or manually editing the CSS code generated by the AI system.
Potential for wider use
But it’s not hard to see how Keyframer could be useful to way more people than designers and animators.
A small business needing social media content, for example, could take a static product shot or other related image and quickly and easily generate eye-catching video content for Instagram or TikTok.
Consumers, too, are likely to find it a fun extension of the way they use photo filters at present.
Both MGIE and Keyframer are at a very early stage right now, so we shouldn’t expect to see them as part of an expected major AI push in “the biggest iOS update ever” later this year – but I can see both technologies making an appearance on Apple devices further down the line.
Photo by Aldi Sigun on Unsplash
FTC: We use income earning auto affiliate links. More.
