When I was about 13 or 14 I decided to teach myself to draw. I’d been getting progressively more interested in anime and manga, and while I’d never shown a huge amount of artistic talent, I wanted to learn. I was moderately successful at it, it took a lot of daily practice but I got better. Eventually I was good enough to sell some of my prints at comic conventions and win an under-18s art competition. A job decently well done.
But the universe didn’t have a career in the arts in its plan for me. I’d improved and was committed, but I always lacked the artistic flair that truly talented artists have. I got a B in Art GCSE, decent. There was always someone slicker, quicker and just better. It’s not enough to be able to clank around with a pencil, becoming a good or great artist requires a depth of understanding and technique that I did not have. I moved on to other pursuits, and my skills atrophied.
With this background it’s been interesting watching the blazing rise of AI art in the past few weeks. AI art generators allow users to type in a number of word prompts and then the algorithm will create an image using them. Prompts can be broad “a green tree” or robustly specific “an astronaut / lounging in a tropical resort in space / in vaporwave style”. The DALL-E program was the first notable AI art generator. Initially released in January 2021 to a select clientele, competitor Midjourney joined in July 2022, while open-source Stable Diffusion released in August 2022. By September 2022 DALL-E opened up usage to anyone and everyone. An avalanche of debate, discussion and galleries upon galleries of AI art ensued, with DALL-E alone now producing 2 million+ images a day of everything from puppies to aliens to anime girls.
Initially any interest in the technology from wider culture was driven by curiosity and novelty. Isn’t it funny that I can put Ronald McDonald in Star Trek, haha. But this initial fallaciousness soured, and battle lines are being drawn. The way AI art generates its pictures is through algorithms of existing art. Partitioning libraries of existing images throughout the web and frankensteining them together at your whim. Your prompts can specify a particular art style or artist’s style in the prompt tags. These artists can be obscure, famous or dead, but if their work is on the internet, it’s fair game to be summoned into the mix at will.
Think about what the central conflict is here. At its most blunt and brutal. It’s between creators, people who have spent in many cases years of hustle, ambition and sweat. Late nights, exams, deadlines, slowly grasping towards financial independence and career success. And on the other side are the passive consumers. People who like the images, are curious about the technology and are often baffled by how easy it is to operate and play with. You just type in a few words and there’s your creation, made just for you. It might be a bit wonky initially, requiring further tweaks and revisions. But it’s yours alone.
On a glib level I don’t think the debate about “is AI art real art?” holds huge significance. At least in the realm of technology uptake and experimentation. Most people don’t know or care what art is. They know it when they see it. Art to the majority is a pretty picture, a nice song, maybe a movie that makes them think a little bit. Every day users can scroll through oceans of fanart and novelty meme content, cute, funny, sexy, shocking. Awarding each a few seconds of attention and a flick of the like/retweet if they’re feeling generous.
Art in the deep philosophical sense, with no art for art’s sake, is abstract. It’s a debate that intentionally has no end point and because of this only appeals to elite subscribers. You already have to elect in, so it’s not going to put many people off. Especially if you’re using it to argue against the promethium fire of a new creative tool.
I’m not in the business of hack futurology or evangelism, but taking the work, the gunk, out of creation is going to change the world as we know it. AI visual images are only the end of the beginning. We’ve had AI music for a while, AI videos are in their infancy, but growing, AI movies next? AI video games?
The accessibility and complexity of these systems are improving daily. How many people have vague ideas in their head that are impractical or difficult to make into reality? How many potter around thinking they have a movie idea, a cool concept for a comic, a fresh music genre blend or video game level? How many would just like to try their hand at iteration, making something based on a franchise or album they love, or a series that never got that second season?
What prevents 99.99% of art from being made is the lack of physical skill which can be applied to put ideas into practice. If you can’t draw, paint, play or act your options are to either learn yourself or pay someone else to do it. If you don’t have any money you’re not going to pay anyone. So you either use your own bad skills and no one likes what you’ve made, or you’re one of the lucky and talented few who can bring your fully fledged form to life in grand technicolor.
George Lucas hit the ultimate creative bullseye making Star Wars. Changing the world forever and managing to convey his own highly specific vision well enough for other people to understand it and fall in love with it. His own talent, the actors, special effects teams, production money and audience all lined up at just the right time. Imagine a future where everyone can generate their own 3, 9, 200 part Star Wars-style sci-fi series. Where The Lord of The Rings never has to end and every dead artist in history can be necromanced from the grave to dance in VR and sing in AI. Because that world is coming.
It sounds like liberation? Your hyper little brother can create a Dinosaur Megazord Pikachu hybrid, you can create the dream image of you and your girlfriend’s future wedding, your aging grandpa can paint a picture of his long-lost childhood home. AI looks to be able to free the vast majority of people who do not have the skills to express their ideas. Even as the technology greedily hoovers up the creative output of those who have put in the sacrifices and work to bring their ideas to life.
I resent the idea of artists as smug skill “havers” while the sullen masses are left as deprived “have nots”. Those skills were acquired with sweat and blood. But what is the answer here? Because AI art looks to me like a steamroller, which is going to crush, or at least mangle beyond recognition, artistic production as we know it.
Something feels different with AI art. Web3, the Metaverse, Crypto, NFTs. Mark Zuckerberg wants to own your brain, Crypto lets you quit your job at McDonalds and become a millionaire. NFTs, with their own promise of fairer rights for artists, but their impenetrable cultish exterior. At the moment these are all bespoke, techie clubhouses which mean a lot to some and nothing to others. And there’s money money money sloshing round in every direction. Promoters and critics battle it out. Fortunes are made and lost. But these technologies all need to be explained. You have to watch a 45-minute lecture to start grasping at crypto trading, let along join in yourself. We’re sinking our costs and learning now to buy in for the promise of tomorrow.
AI art does not need to be explained, it exists right in front of you. This is why the explosion is happening in a way it hasn’t for other technologies. This is an iPhone moment, you don’t need to tell me the benefits- I can see them. Art is a pretty picture, AI art can produce for you a bespoke pretty picture, and those pictures will keep getting prettier and prettier. A kitten, an old bike, a castle, a beautiful girl in a sun dress smiling at you and only you, a seaside view, a Panzer tank, a rabbit, Ronald McDonald. It’s right there for you. Whatever debate is to be had is had after the image is made.
In 2000 Lars Ulrich, drummer of iconic thrash metal band Metallica attempted to sue fans who had downloaded the band’s music illegally from the nascent Napster. Now seen as a pointless and damaging act of Boomer idiocy, which made Ulrich aggressive and profoundly uncool, Ulrich was still prescient in his understanding of the situation. If not in how he chose to express his concerns.
Music piracy (even calling it that now seems so antiquated)swallowed everything and the music industry that Ulrich was attempting to defend was brought to heel by the end of the decade. Consumers wanted more for less, simple as. No pleading, threats or corny PSAs were going to stop this. You could sue one person and a thousand would take their place. The pleasure derived from free music was always going to win.
AI art will become like music piracy, there is no way to stop the drive of pleasure. Being able to create your own images of whatever you desire or need is unlikely to be tut-tutted away or argued against. Not because the legal and moral arguments of artists don’t hold weight, they do, immense weight, but because the change is too fundamental to be stopped. This is not even beginning to unpack how AI will transform the wider workplace and economy. The floor is shifting under all of us very quickly.
I have every faith that the talents of artists will find new avenues for creation and success after the AI revolution crashes into us all. But AI is coming for everything and everyone, and will utterly remake society into images we can scarcely imagine.
All images were created with a web version of Stable Diffusion in a very rudimentary fashion, they’re not that good. But they’re not bad either.