Apple's Latest AI Tool Can Animate an Image Based on Your Description

[MacRumors]

Apple has made another addition to its growing AI repertoire with the creation of a tool that leverages large language models (LLMs) to animate static images based on a user's text prompts.

MacRumors image made with DALL·E​

Apple describes the innovation in a new research paper titled "Keyframer: Empowering Animation Design Using Large Language Models."

"While one-shot prompting interfaces are common in commercial text-to-image systems like Dall·E and Midjourney, we argue that animations require a more complex set of user considerations, such as timing and coordination, that are difficult to fully specify in a single prompt—thus, alternative approaches that enable users to iteratively construct and refine generated designs may be needed especially for animations.

"We combined emerging design principles for language-based prompting of design artifacts with code-generation capabilities of LLMs to build a new AI-powered animation tool called Keyframer. With Keyframer, users can create animated illustrations from static 2D images via natural language prompting. Using GPT-4 3, Keyframer generates CSS animation code to animate an input Scalable Vector Graphic (SVG)."

To create an animation, the user uploads an SVG image – of a space rocket, say – then types in a prompt like "generate three designs where the sky fades into different colors and the stars twinkle." Keyframer then generates CSS code for the animation, and the user can then refine it by editing the code directly or by entering additional text prompts.

"Keyframer enabled users to iteratively refine their designs through sequential prompting, rather than having to consider their entire design upfront," explain the authors. "Through this work, we hope to inspire future animation design tools that combine the powerful generative capabilities of LLMs to expedite design prototyping with dynamic editors that enable creators to maintain creative control."

According to the paper, the research was informed by interviews with professional animation designers and engineers. "I think this was much faster than a lot of things I've done," said one of the study participants quoted in the paper. "I think doing something like this before would have just taken hours to do."

The innovation is just the latest in a series of AI breakthroughs by Apple. Last week, Apple researchers released an AI model that harnesses the power of multimodal LLMs to perform pixel-level edits on images.

In late December, Apple also revealed that it had made strides in deploying LLMs on iPhones and other Apple devices with limited memory by inventing an innovative flash memory utilization technique.

Both The Information and analyst Jeff Pu have said that Apple will have some kind of generative AI feature available on the ‌iPhone‌ and iPad later in the year, when iOS 18 is released. The next version of Apple's mobile software is said to include an enhanced version of Siri with ChatGPT-like generative AI functionality, and has the potential to be the "biggest" update in the iPhone's history, according to Bloomberg reporter Mark Gurman.

(Via VentureBeat.)

Article Link: Apple's Latest AI Tool Can Animate an Image Based on Your Description

 

[swiftapplefan]

That’s actually really cool, can’t wait to see how they’ll implement it in their products.

 

[jayryco]

I think it's cool but the implications of these tools on the creative industry will be devastating once brands and corporations find the feasibility of an AI far cheaper than creative professionals. Especially in the next 5-10 years.

 

[purplerainpurplerain]

I think it's cool but the implications of these tools on the creative industry will be devastating once brands and corporations find the feasibility of an AI far cheaper than creative professionals. Especially in the next 5-10 years.

Incorrect and badly ignorant. Don’t listen to stupid Wall Street AI bros.

Companies employ people to use the tools. They can’t just have some executive sitting there asking AI. The executives and managers have other things to do.

Companies require their creative departments to use the tools.

The tools like this are very imperfect and unpredictable. They always will be. You cannot replace all traditional workflows with ‘talk to AI’. That can only happen if you want everything to be really sht.

The output of AI tools is combined with other materials that are all made normally. The creative professionals do this.

 

[Thomas Davie]

Now I can create an endless number of Beavis & Butthead episodes; just what expensive consumer computers were always meant to.

There are *some* really good ai music creations. Post Jim Morrisson, the first Doors album was ‘Other Voices’ with a stunning song; Ships with Sails. I always thought that cried out for a Jim Morrisson rendition, but he died a little bit too soon. well, someone has gone and created a version of the song with Morrisson on lead vocals.

Some stuff is pretty weird, like Frank Sinatra singing ’The End’ and others, like the Morrisson/Ships example are also very good, such as Ronnie James Dio singing Traveller in Time by Uriah Heep (thatwould have been a concert I would pay to see). But yeah; Beavis & Butthead teach cavemen how to fire pr Cornholio encounters a national TP shortage.

Oh the glorious possibilities.

Tom

 

[purplerainpurplerain]

Now I can create an endless number of Beavis & Butthead episodes; just what expensive consumer computers were always meant to.

You’ll get sued by Beavis. Issue fixed.

 

[Thomas Davie]

You’ll get sued by Beavis. Issue fixed.

Nah, won’t happen, b/c I’ll eat a lot of sugar and turn invisible.

Tom

 

[IllegitimateValor]

How many iterations before the AppleAI in Vision Pro can give us Spatial Computing versions of hallucinations while we’re still sober and lucid?

 

[LV426]

When I can get Max Headroom on my iPhone instead of Siri, I will be happy. Remember, kids, two’s company; three’s an audience.

 

[Rafael Santos]

How many iterations before the AppleAI in Vision Pro can give us Spatial Computing versions of hallucinations while we’re still sober and lucid?

That's a use case I'd be interested in.

 

[Rradcircless]

That Apple sell Logic and Final Cut, widely used engines of human creativity, I really, really, really hoped they'd have had the perspective that human art is sacred.

I thought they could angle toward AI's system integration and offer things such as recognising user behaviour and adapting performance or recommendations, rephrasing and reorganising notes, more intelligent file searching like finding things that match a description (I'm studying and it's so hard finding the schoolwork that matches what I remember that I go online and find a new source every time).

 

[IllegitimateValor]

I think it's cool but the implications of these tools on the creative industry will be devastating once brands and corporations find the feasibility of an AI far cheaper than creative professionals. Especially in the next 5-10 years.

The two chief dangers to worldwide society I see in some future iteration of AI/LLM image technology:
- people with sufficient quantities of money (to publish or platform content) no longer needing to hire humans for creating media, impoverishing large swaths of people and continuing the distortion of culture towards the views of those with power

- news being able to be faked at such scale that discerning what is true becomes impossible (to the benefit of maintaining status quo for those with means)

 

[Zest28]

That Apple sell Logic and Final Cut, widely used engines of human creativity, I really, really, really hoped they'd have had the perspective that human art is sacred.

I thought they could angle toward AI's system integration and offer things such as recognising user behaviour and adapting performance or recommendations, rephrasing and reorganising notes, more intelligent file searching like finding things that match a description (I'm studying and it's so hard finding the schoolwork that matches what I remember that I go online and find a new source every time).

Science has shown that music is all the same in the end. People copy each other and call it “inspiration“, which is what AI can do without any problem.

AI is a good thing, because it forces people to be actually creative for once instead of creating the same thing people did before them, or else be replaced.

 

[Wanted797]

I think it's cool but the implications of these tools on the creative industry will be devastating once brands and corporations find the feasibility of an AI far cheaper than creative professionals. Especially in the next 5-10 years.

AI won’t take our jobs. People who know how to use AI will.

 

[Rradcircless]

news being able to be faked at such scale that discerning what is true becomes impossible (to the benefit of maintaining status quo for those with means)

Oh yes, and to manipulate images per the previous Apple AI development, your regular Joe friends on Facebook could fake anything like making their homemade meal appear better than it is, women with body dysmorphia could synthesise the idealised physique... Reality and history, both on an individual and global scale, will become a canvas to freely paint over, leaving no fossils to rediscover the 21st century in the far future.

 

[Rradcircless]

Science has shown that music is all the same in the end. People copy each other and call it “inspiration“, which is what AI can do without any problem.

AI is a good thing, because it forces people to be actually creative for once instead of creating the same thing people did before them, or else be replaced.

Then why doesn't every artist get sued for copying those before them? Why is it that AI is stealing static art (paintings and whatnot) but not music? Because the music industry would sue them to oblivion. It may all be derivative, and it's human. What artistic value does AI offer anyone other than as a collaborative tool used by the true human artists?

 

[Zest28]

Then why doesn't every artist get sued for copying those before them? Why is it that AI is stealing static art (paintings and whatnot) but not music? Because the music industry would sue them to oblivion. It may all be derivative, and it's human. What artistic value does AI offer anyone other than as a collaborative tool used by the true human artists?

Just watch, AI is going to make better music than your “human art“, because all you do is making the same crap, which is easily replaceable by AI.

Only real artists will survive that are actually unique and not some derivative of other people their work, which will be a very very small minority.

People who worry about AI are people who are not real artists in the first place.

 

[krspkbl]

Are we meant to be impressed? It's good they are catching up but we've had AI image generators for years at this point.

 

[DTphonehome]

Hey siri, can I get a hat wobble?

 

[HeinoM]

I think it's cool but the implications of these tools on the creative industry will be devastating once brands and corporations find the feasibility of an AI far cheaper than creative professionals. Especially in the next 5-10 years.

I had the exact same thought, only I fear it may happen even sooner than that. Major publications are already laying off writers in droves as they transition to AI generated content. Creative careers of all types are going to be devastated.

 

[Hardijs]

How many iterations before the AppleAI in Vision Pro can give us Spatial Computing versions of hallucinations while we’re still sober and lucid?

y'all need a jailbroken device for that. EU buddies will get that in alt stores

 

[victorvictoria]

I have seen the future, and am happy I won't be around to witness it.

 

[Hardijs]

I had the exact same thought, only I fear it may happen even sooner than that. Major publications are already laying off writers in droves as they transition to AI generated content. Creative careers of all types are going to be devastated.

some super creatives / pol agenda aligned ones will be retained for the time being. the other 80% definitely will be gone in 3 years.

 

[tehabe]

I hope this generative AI/LLM trend is ending sooner than later and all those companies go bankrupt.

 

[return2sendai]

Science has shown that music is all the same in the end. People copy each other and call it “inspiration“, which is what AI can do without any problem.

AI is a good thing, because it forces people to be actually creative for once instead of creating the same thing people did before them, or else be replaced.

Gotta say, despite someone’s angry face, I wholeheartedly agree. I’m a hobbyist writer and what AI produces is utter rhubarb compared to my modest scribblings. Until AI is a sentient, perceptive, emotive, intellectual, multidimensional super-being, it will never hold a candle to this sentient, perceptive, emotive, intellectual, multidimensional super-being.