How to Embrace Artificial Intelligence to Make Better Animations
Philip K. Dick once wrote a novel pondering whether androids dream of electric sheep. Fast Forward 54 years and artificial intelligence (A.I.)—defined by Oxford as “computer systems able to perform tasks that normally require human intelligence”—has come a long way. We may not know what they dream of, but to paraphrase another PKD title, AI’s can now visualize our dreams for us, wholesale.
Incredibly sophisticated, though not quite yet sentient—depending on who you choose to believe—algorithms that spent years stealthily learning to interpret and visualize the intention of our words, thereby mimicking the technique of traditional artists, have arrived.
Specifically, a spate of deep learning AI visualization platforms hit the creative scene like an asteroid in 2022. These innovative systems invite you to feed them prompts—phrases describing an image or a scene—then spit out small selections of pictures attempting to match the vision you’re holding in your mind. No doubt you’ve seen the otherworldly portraits, epic landscapes and impossible macro shots sweeping over Instagram. Heck, we first caught sight of these cybernetic newcomers on LinkedIn of all places (thanks to Nando Costa’s wonderful thematic explorations).
These systems are developing at an unprecedented pace and they’re dragging the entire creative advertising industry with them on their roller coaster ride. Cosmopolitan published AI’s “first magazine cover” in June of 2022.
Rapid-fire changes to this nascent technology mean it’s impossible to predict its future though. There’s no telling, for instance, how it might impact the ways we produce animation and motion graphics in the years to come. Anyone who tells you they know exactly what’s next is, frankly, full of it.
What we do know is that AI offerings are forcing the first real paradigm shift to hit the creative scene since the onset of Adobe. Spoiler alert: it’s kind of a BFD; and those of us at MOWE, are excited for the opportunities and challenges this new technology will usher in.
Still, we wouldn’t fault you for sporting a little anxiety: “How will AI impact my job? My workflows? What does this mean for how we produce content?” Before your wonder turns to worry, let’s dive into common AI-related questions we can try to answer—those related to the wonderful world of animation.
Will Artificial Intelligence Replace Your Animation Team?
No. AI is a valuable tool with the potential to streamline work, which is why we’re committed to staying on top of every AI iteration. In its current form, it’s a bit like adding a new teammate to the studio: a really fast research assistant that can help directors build more effective and cohesive mood boards.
Only it’s not a replacement for human animators and their deep understanding of the arts and visual storytelling techniques. Not to mention the can’t-put-your-finger-on-it quality that makes artists, well, artists. That can never be replicated. And don’t forget the prompts. Modern AI requires a human operator to plan and describe desired scenes, not unlike how a camera operator or a focus puller coax poetic imagery out of their own imaging machines.
At best, there may come a day far down the line where AI directors might compete with human directors—but such competition is nothing new when producers already spend much of their days sifting through hundreds of reels in search of the perfect animation partners for each project.
We also believe there will always be something to be said for working with partners you trust. It’s no secret that human relationships are the cornerstone that makes great work possible.
How Did We Get Here?
Image making AI’s have been around for years. But in spring 2022 startups like Dall-E 2 and MidJourney emerged to promote low-res results that saturated social media. This new generation of AI-driven images looks less dreamy and fractal and more like actual serviceable art.
Industry leaders took note. The paradigm shift began. And the images have gotten more sophisticated nearly every day since (literally). What began as still frames has evolved into trippy movements with motion generators such as DeForum and DiscoDiffusion. And by the time you’re reading this something even newer and more exciting will have hit the scene.
What Are The Strengths of AI for Animation?
There will come a day—sooner than later—where AI will be integrated in all our creative products. Just like any other tool, AI has its strengths and weaknesses (see below for limitations). Its main strength is undoubtedly an uncanny ability to help artists prototype proof-of-concepts faster.
Consider this: each new project starts with a mood board or a director’s treatment. Studios create these by sketching new art, collecting old references, or more often than not, scouring sites like Dribbble, Behance, Colossal, Instagram, etc. for handfuls of images that approximate the style we’d like to pursue. This allows us to communicate our goals and ensure that we’re speaking the same creative language as our clients. Only it can take days, or longer, to develop and refine the perfect mood boards.
AI has given us the ability to reduce some of this time, making it possible to quickly adjust concepts, develop rough images on the fly, and hone the reference imagery we’re using to closely align with a brand’s voice and color more than any combination of found images might have previously allowed. That is to say, AI enables us to communicate more effectively and help our clients get a better view of the end result before we really begin working in earnest.
We used MidJourney this summer to supplement a data visualization mood board on a quick turnaround project for The Trade Desk. The results weren’t perfect, but they helped us paint a vision of where we saw the spot going.
The AI images we shared sparked conversations around motion graphics behaviors and the meaning behind each of the datatypes we ultimately designed for the video. The brand ended up choosing to pivot away from those initial images, but they still helped us quickly find our way to the stylistic sweet spot that worked for the brand, which we proceeded to design and refine by hand and then vector.
There is an element of risk involved. AI won’t always give you what you want, and that’s what makes it special. Because identifying what one does not like, and learning to understand why that is, can be just as informative, if not moreso, than trying to match what one likes when crafting an animation—or any other creative product for that matter.
What we can definitively say is that AI gives art directors the freedom to “fail fast” in developing new styles and concept designs. If an AI generated image hits the mark, great! We can move it to the next step of development and refine with Adobe. Or we can trash “failing” images that miss the goal completely and tweak the prompt to aim closer to our target and run it again. AI visualization is a game of trial and error and correction. It’s not flawless by any means, but it’s a nice option to have for the times you need to generate lots of complex prototypes in a short amount of time, something that might otherwise be too costly or risky without the support of AI.
In sum, embracing AI has the power to bolster communication top-to-bottom between animators, agencies, brands, studios, etc.. It can break down barriers to the free sharing of ideas and constructive collaboration. And arriving at a final animation deliverable — or even an aesthetic for a future animation — is simply easier when the creative director, the producer and the director are all able to confidently say they understand each other’s intentions.
What Are AI’s Motion Graphics Limitations?
AI is a promising new starting point — like a rocket booster getting us into orbit. But just as rockets depend on control centers full of (human) engineers, animation will continue to rely on artists and strategists with the expertise to guide the execution of a story.
Put simply: The vision of an artist is irreplaceable. That’s the main weakness of AI. After all, what phrase would you enter into an AI image generator without an artist’s fledgling idea?
Social media posts make some of the results look flawless, but in reality, the current output of these platforms is still rough upon close inspection. Images are low resolution, grainy, missing odds and ends, or downright bizarre when rendered at higher resolutions. AI doesn’t currently allow you to revise images with enough precision to arrive at a finished piece. In fact, the most impressive AI results are the ones artists remove from the system and “paint over,” leveraging the AI image like a pencil sketch they layer with color.
You also can’t replicate the same results consistently with current AI platforms. Even if you type in a phrase verbatim, you won’t get the same results twice. That makes refinement within the system difficult.
Keep in mind, these are still, essentially, first gen technologies, and those of us dabbling in AI today are the early adopters.
Sidenote: Even the legality of who owns AI-generated imagery is still up in the air. Get to know the latest AI Bill of Rights and read the fine print before you start generating images. Unresolved legal questions, such as where AI platforms source their imagery, are another weakness of AI as it stands today.
AI Will Spark the Next Golden Age of Creativity
Some professionals out there may be concerned that AI platforms will stifle originality. Or change the way artists work for the worse. On the contrary, we think AI will jumpstart a renaissance of creativity — regardless of its noted strengths and weaknesses as a design tool.
Why?
Because AI is conformative by nature. It’s (pun intended) robotic. And artists resist conformity. Period. Flat design of the past decade unseated skeuomorphism in the realm of user interfaces the same way that cubism followed closely on the heels of impressionism. Juan Miró even famously burned his own paintings in response to society’s demanding expectations of his art to conform to styles he had previously explored. It’s an intrinsic trait of the artistic spirit that brands rely on to create newfangled content that stands out in the crowded marketplace. AI also gifts artists with diverse imagery that can serve as unique fodder for creativity.
All to say, rather than dampen artists’ imaginations and creativity, AI is likely to (unintentionally) ignite them.
The AI-Related Questions Have Only Just Begun
We’re not worried AI will replace humans. We don’t see a future where animation is 100% robot-driven. Our biggest concern is remaining at the forefront of what’s next for AI.
We’re witnessing a Moore’s Law-type of growth in AI technology, only AI platforms seem to be making generational leaps every few months rather than every two years. And the only people who can even attempt to predict what’s next are the developers guiding AI’s growth, which is largely done behind closed doors in the secretive “stealth” modes favored by silicon valley investors.
The best way to stay in the know as of now is to engage the wider AI community. MOWE is committed to doing so by providing feedback to developers and continuing to test AI platforms. And of course, we’ll keep you updated on ever-changing AI technologies in the process.
As to Philip K Dick’s probing of the existential barrier between human sentience and artificial intelligence, we prompted: “do androids dream of electric sheep?”
We’ll leave you to ponder the meaning behind MidJourney’s response.