Why Did AI Art Explode in 2022?

Midjourney, DALL-E 2 and other creative AI platforms have either excited or scared the living hell out of creatives in recent months. The ramifications of their impact have been hotly debated. There are those that have dived deep into exploring its possibilities, which let’s be honest are impressive but require a good grasp of coming up with good word prompts that produce the desired result. And then there are others that are terrified, like illustrators and concept artists who think that their careers are doomed.

 

Let’s be honest here… Typing random stuff into one of these tools and seeing what it spits back out is rather entertaining. It is new. People have never seen it before, well not in the way that it is in the palm of their hands and is near enough immediate. Which is one of the reasons it has exploded so much. It is one of the first visible examples of the power of AI that is for the masses. It is also an example of how AI can express creation or creativity. Something inherently not associated with AI. Being creative was supposed to be the one thing AI could not do.

 

But is it ‘really’ being creative? I would suggest it is not, but instead, it is just another tool for creative people and artists to use. Some will embrace it and others won’t, but those that do it well and harness its power will both speed up their workflow and be in demand, looking to combine it with their current tools and styles to create new work. I think in the long term this could lead to clients or agencies wanting a better quality of work, faster and cheaper. Which leads to other challenges… How do you know if someone is using it or not in their work? Which in turn leads to questions around ownership, usage rights and copyright. A can of worms is indeed being opened.

There’s no doubt that AI-generated artwork made a huge impact on the art and design scene this past year. Midjourney brought millions of users to Discord with most just playing around with it while others were legitimately looking for a tool to generate their vision. There was, and still is, a lot of trepidation with using AI in a professional context. 

It comes down to ownership, which in my mind, really comes down to how that AI was trained to generate the images in the first place. If you look at what’s behind the code, every deep-learning algorithm is fed a litany of images. Most of those images come from online sources; images that were captured and created by artists. That’s the ethical line we, as an industry, need to reconcile. 

Although the images DALL-E, Stable Diffusion, and Midjouney are making are wholly unique to themselves, they ‘trained’ their artificial eye on artwork human beings have created in the past. But don’t we all reference each other’s work? It gets tricky, doesn’t it? Let’s consider a different approach here. Treat AI not as a replacement to our pipelines, but as another tool in our toolbox, we can use to better our craft. We’re already using it as a tool for reference, but let’s use it deeper in the pipeline and feed those deep-learning algorithms with original art and design and allow it to do what it does best: iterate. 

Use AI to generate hundreds of deliverables in a social campaign based on an original set of designs. It’s the difference between letting the AI come up with the look and the feel versus a human hand creating a new look no one has ever seen before and letting the AI generate iterations of a theme.

For the creative mind, embracing AI is like the move from DVDs to streaming - you know that ultimately it’s a good idea and will help you to do more, faster - but it’s hard to avoid being concerned about losing that very human instinct to be in control. We want stuff in our own hands to know it’s going to work, especially if our brain is wired creatively. However, like all things tech, it’s about jumping in feet-first. 

So why is it taking off now? It’s a natural step forward in our tech-driven world and the creative world needs to find a way to accept it and work with it rather than against it. Learning to be comfortably uncomfortable with AI could lead us to more original, unique work – it’s another tool in our arsenal, already all around us (just think about your smartphone) and, if we don’t embrace it, we risk being left behind.

The emergence of AI tools for creativity in the last six months is akin to the move in the early ‘90s from hot glue paste-up to desktop publishing. Many of today’s traditional production capabilities - confined to a highly trained production staff - are now available to anyone simply by using natural language prompts. An era of desktop production is upon us. And technology is advancing at rates unparalleled by our past experiences.

As of this moment, we now can produce any still image we can imagine without needing a photoshoot by using natural human language. By year’s end, this capability will extend to production code, animation, 3D objects, 3D environments, music, voice narration, film footage and video games. All of this is generateable with tools that are no more difficult to use than email.  

It will change the work we do. The way we do it. How we staff. Who we hire. How we train and what we charge. It is a seismic shift and likely one of the most underrated stories of the decade.

As a Groupe, we’ve been tracking these developments closely. We’ve used these technologies to pitch and win new business, prototype in real-time and present clients with new offerings. Through newly formed partnerships, we are beginning to roll out new capabilities to a small set of early adopters in our agencies globally.

Ai is not a fad. It will be embedded in almost every tool we use and it will challenge the notion of who and what we consider creative. It may break some traditional agencies, but it will launch a thousand more.

AI art is improving at an astonishing rate. The images are getting more and more natural. The downside of this is the way it works is changing as well and you have to keep relearning prompts and settings. It’s another important tool all creative people should learn. And it’s not just image generators - AI is improving strategy, copywriting and more.  

AI image generation is already changing the creative process. In addition to comps, we're now using it for consumer facing images. This opens up a legality and usage debate which I’m sure will evolve with the technology.

The bottom line is, AI is an extremely powerful tool, just like Photoshop. It’s not a replacement for ideas, wisdom and craft but it’s a facilitator, a power tool for creative (and non-creative) humans.

The only constant in the world is that it's changing. Since the dawn of time, artists have used and adapted various tools and techniques, from sticks and stones to pen and paper, coloured water and oil, chisel and hammer, and digital software to realise their vision. Ultimately, the heart of art is the vision of the artist, whichever tool she decides to use.

Text-to-image-based natural language platforms like Midjourney, Dall-E, and Stable Diffusion, allow more people to turn their artistic vision from ideas into reality, so the net effect here is positive. The more accurately these models follow the given descriptions, or ‘prompts’, the more control the artists have over the result to reflect their vision making their imagination the key to unique results. So, in the end, compelling art is still about original ideas, not the tools you use to make them.

Sure, these new tools offer a shortcut to technical excellence, but like in filmmaking, what the image is about plays a more vital part in creating unique and memorable art than technical excellence. Even though modern technology makes it possible to conjure anything you can imagine onto the big screen, it's still the story that matters when making a great film. The same applies to still imagery, regardless of how we create it.

There's also a discussion about using existing imagery to train these algorithms. Regardless of the tools or medium, all art is inspired by the works of artists who came before. Unless you grow up in a vacuum, you absorb the world around you and get inspired by the art you see, so when you produce your own, it's influenced by what you've experienced before. Then you put in your own spin that will again inspire someone else. And so it goes into eternity.

Many artists have already embraced these new tools adding them to their ever-growing toolbox, and the best ones are still the ones with the best ideas. I can't wait for these tools to evolve from text-based to spoken-word models, which is inevitable.

When text-to-image AI programs blew up this summer, we saw an opportunity to reinforce Heinz’s status as the world’s favourite ketchup. While everyone was posting weird and hilarious AI-generated images like ‘Napoleon riding a Harley Davidson’ or ‘R2D2 getting baptized’, we wanted to use these new platforms to reinforce to a younger audience that Heinz is the definitive ketchup. 

We turned to DALL-E 2, the most advanced Al image generator - so advanced that it wasn’t even available to the general public yet - and asked it what ‘ketchup’ looked like. The result? Even to Al, ketchup looks like Heinz. And as the prompts got weirder, from ‘Renaissance Ketchup Bottle’ to ‘Ketchup Tarot Card’, the Al still generated results that looked like Heinz. 

Then we asked people to push the experiment further with suggestions for new ketchup-based image prompts. We took these suggestions and turned them into the first-ever campaign with visuals generated entirely by AI. 

For us, the most exciting aspect of the campaign was the fact that people could submit their own ketchup-based prompts and see them brought to life in vivid detail. That’s a big part of why this technology has blown up. In a way, it democratised art. Anyone could have an idea, no matter how unusual it was, and have it beautifully rendered within seconds. Then they got to share that idea with the world. 

As for the role that text-to-image programs will play in the advertising industry, we still think creativity is the most powerful tool to drive brands forward. This is another tool to push our creative capabilities forward and execute ideas that may not have been possible before. Artificial intelligence will help bring ideas to life, but the creativity of people will always be the most powerful tool we have at our disposal.