If life is a party, then AI is the hottest place to be right now for anyone involved with creative services, whether you’re a designer, an artist, an engineer, a researcher, a producer, an editor, or a product manager.
I’m using a party analogy because dancing with AI is FUN. It makes me feel like a kid again. And it’s not just me — it seems like in the last month, everyone I know who is obsessed with visual expression can’t get enough of AI-driven art. We’re all lining up to get invites or early access to the most exclusive clubs, like DALL·E 2 (Powered by OpenAI, a competitor of Google’s Deep Mind) and MidJourney (now in open beta).
Of course, we’re grown-ups and people who make our living in graphic or digital communication. And we see the power of AI to get the random sparks of brilliance quickly out of our heads and into real life. For me, it’s like I’ve got the power of an entire concept team in my pocket.
Rocking my baby asleep for a nap, and inspiration hits? I pop open Discord on my phone, fire up MidJourney, and fire off the concept:
Even if I’m not divinely inspired, I often use AI tools to relax or keep my creativity flowing between meetings. Perfecting the art of prompt crafting is truly the perfect antidote to feeling drained and sick from doomscrolling on TikTok, Twitter, Insta, or any other social channel.
This isn’t to say everything about AI is hearts ❤️, flowers 🌸, rainbows 🌈, and unicorns 🦄. AI has many dark sides, as it reflects society — and things can be pretty grim these days. I could do a whole separate article about how things like skewed datasets promote bias, malignant stereotypes, pornography, intellectual property stealing, and a myriad of other dark facets — but, like I just said, that’s a whole other article.
Instead, this article includes my experience of spending countless hours enjoying this AI design shindig, which has permanently shifted how I work, play, and live — for good. And the dance with AI is specific; you’ve got to lead. Otherwise, you just may render yourself obsolete. Because things have changed rapidly in the last year, and from where I sit, I believe AI will only keep improving — even as it profoundly changes our world.
So, let’s get this party started and see what makes AI such a fun wingman or wing woman or, probably most accurately, wing-thing. (And how to keep it that way 😉)
A Friendly Get-Together
First, a quick primer on how AI-driven art and image creation works. A couple of party metaphors for AI back up my assertion that the technology will only keep getting better and better from here.
Let’s start with a concept from nature: biomimicry, which is essentially design perfection. Biological organisms develop strategies to survive by optimizing their organization and functioning. Then, they adapt their form to the function.
In other words, humans (and other living creatures) evolve to fit into the most desirable, most happening scene.
AI-driven artwork is typically created using a Generative Adversarial Network (GAN). Essentially, it’s a deep neural network framework that learns from training data and makes new data (i.e., an image) with the same attributes as the training data. It does this through two competing neural networks — a generator (the network that learns to create realistic fake data) and a discriminator (the network that judges the data and deems it fake or not).
The discriminator improves the more bad fakes it sees, and the generator improves by learning how to make it past the discriminator’s tough standards.
In other words, the generator is trying to get into the club with a fake ID, and the discriminator is there to be a bouncer and reject crappy fakes.
The ongoing struggle between the two is what makes AI art continuously evolve. And the way we know both are fighting desperately for dominance — and both winning — is that finally, in the last six months to a year, AI-created art has broken through to the mainstream.
How AI Became the Belle of the Creativity Ball
It’s been seven years (2015) since a group of researchers questioned if AI-driven automated image captioning (then a novelty) would work in reverse: could AI turn text into images? And not just any graphics — incredible images that people had never seen before and don’t exist in the real world.
While the technology initially turned out blurry blobs, today, it’s become much more precise. And it’s opened up a veritable treasure chest of tools — not just exclusive ones like DALL·E 2 and MidJourney, but also free, fully open tools, including Craiyon (by Huggingface.co) and Disco Diffusion, which make it easy for engineers to spin up their own applications and tackle similar creative feats.
Thanks to GANs, the more AI is trained on text and images, the more “imaginative” and evolved it becomes. Of course, the tech’s version of imagination is mathematically driven, whereas human creativity is a more unique, organic process, I’d argue. I believe our imagination will always be the OG source code. (Well, at least for the foreseeable future, say the next few decades.)
AI and machine learning aren’t only taking faux image creation generation by storm; the technology has also become a mainstay in optimizing real-life images. Nowadays, we all travel with high-powered cameras in our pocket, and what gives them a leg-up over the typical lens- and sensor-driven cameras is the software. AI-driven computational photography allows phone cameras to optimize images (remove blurriness, fix lighting) with no post-production necessary — you can literally use the pictures for professional purposes without needing to know how to use professional tools like Adobe Lightroom/Photoshop. And then the whole library becomes organized so we humans can take better advantage of the photos we shoot.
I use this example to say, fearing AI and refusing to use it (on the grounds that it’s something that’ll replace you) is silly. And it’s already too late. AI is part of our day-to-day lives.
So, establish your position on the dance floor. Let AI minimize the mundane while you maximize the space to create and spread your magic far and wide.
Let’s say, for example, I’m designing an ad campaign for a coffee company. Here’s how the film script starts:
We open on a tight shot of a steaming cup of coffee in focus with depth of field.
Typically, I’d have to pull one of my illustrators away from doing more meaningful work like creating new character ideas or exploring a fresh visual style to have them spend hours working with a writer or director to sketch out storyboards. This is not a great use of their time.
I’d much rather start with my AI sketch and task my storyboard artists with higher-level work. What are the nuances they can bring to the stories we’re telling? What more can they dream up with the time freed up from busy work? And how can they help train the AI to higher levels of innovation and creativity?
The New Dance for Humans
This leads me to what makes me so excited about AI as a co-creator for designers and makers. Instead of fretting about being replaced, we need to start focusing on what the technology really replaces: wasted time.
AI can take on mundane tasks so we humans can focus on using our brain for what it can uniquely do: Solve problems, connect dots, and be creative. That doesn’t limit your earning potential or undermine your job security — it maximizes it.
Here are just a few ways I’ve seen AI-driven art help power workplace transformation, all of which are happening as you read this article:
- My friend who leads the Creative Team at a AAA game design studio uses a commercial license of MidJourney with his concept artists to rapidly create new configurations and ideas. Most people think about AI as an end product, but my friend and his team use it to spark new ideas as a primary tool in their toolbox for character design, environmental design, and level ideas.
- Years ago, Airbnb created a tool like this to speed up their time from wireframe to implementation. This is the same principle that’s fueling the no-code revolution — the dramatic shortening of time from concept, customer input, and release or iteration. This is mission-critical in today’s uber-competitive marketplace.
- Cosmopolitan magazine just used DALL·E 2 to create its cover in 20 seconds. The magazine used the opportunity to represent what they hope will be the case: more opportunities for women in tech to “explore how AI tools can enhance their work and their lives.”
And that brings me to an important point, as the journalist who wrote the cover story, Gloria Liu, perfectly stated:
“My initial takeaway is that DALL·E truly is an artist’s tool — one that can’t create without the artist. Which might ultimately be the point.”
I believe we have to change how we think about value. Is our valuable time worth wasting on tasks AI can do for us? AI art will get faster and better — unlocking many new possibilities for creators and makers. As more sophisticated text-to-image generators like DALL•E 2 become more open and accessible, the use cases will evolve and expand.
As I always say, the tools we make, make us. The party’s just getting started. So, what’s your hot take on AI-art and AI platforms?