Long gone are the days when you could only imagine what Harrison Ford's Indiana Jones would look like as a baby by dressing up your own toddler for Halloween. Thanks to AI-powered software, such as Midjourney and DALL-E, which has helped countless artists realize their wildest ideas with a few clicks of a mouse and hand-calibrated text prompts, we can see The Simpsons' characters as real human beings and a plethora of selfies by iconic figures.
With so many new images being generated every single day - to the tune of 20 million, according to WIRED - it begs the question why so many folks are so hyped up about playing around with AI these days? Well, it's simple, really: these text-to-image canvases of unlimited potential are churning out some seriously impressive things that's just begging to be admired.
The best part? These AI-generated images are totally unique. You won't find one that's an exact copy of the other (although, there's always a slight chance). They're constantly surprising us with new and unexpected patterns, also additional pairs of fingers, never seen before. To steal a phrase from WIRED's Kevin Kelly: "It’s a similar pleasure to exploring a video game world, or paging through an art book." We couldn't say it any better, even with the help of AI.
Of course, nowadays you can't spell (AI-generated) "Art" without some kind of controversy, it appears. Earlier this month, when a German photographer named Boris Eldagsen snatched the first place in the prestigious Sony World Photo Award contest, many were furious because his prizewinning photo was created with the help of 1s and 0s. Or DALL-E 2, to be precise. Thus becoming the first AI-generated photograph to win the award (which Eldagsen later refused to accept).
A similar uproar happened when in 2022 designer Jason Allen won Colorado State Fair's fine art competition with what appeared to be Midjourney-generated artwork enlarged and printed on canvas. It was the early days of AI-powered art making, and so these sorts of announcements took the world by surprise, splitting the public in two.
Milan Jaram, the digital artist known for more traditional digital art, and more recently for his AI experimentations, like the eerily realistic Simpsons' series, has a thing or two to say about the matter. "I do not call myself an AI artist and I do not use 'artist' or 'art' hashtags - I steer away from that controversy personally," he explained to Bored Panda via email. "This is because I am a traditional artist. In fact, my Instagram verification comes from being a professional artist since the '90s."
Like many traditional artists whose fundamental skills are deeply rooted in hand-drawn artwork, Milan believes that there's nothing wrong with the harmony between AI-generated stuff and traditional artistry. It just, so to speak, needs a little time and calibration - just like generating pictures via Midjourney and DALL-E.
"Remember, 'art' is subjective. A good example is an AI artist booth next to a traditional artist booth. If someone is going to buy the traditional artist's art, they will - whether the AI artist booth exists or not," explained Jaram."People are buying your art because they like your style. They put a value on it. In this scenario, there won't be very many people in an AI artist booth, and if there will - prices won't be high. The freedom to create via AI muddles any value it has."
To see what Milan means, one has to turn his/her/their heads to Mike Winkelmann, who's more widely known as 'Beeple'. A few years back, this digital wizard became part of art history books by selling his NFT art collection for an eye-watering sum of $69 million. And if only you didn't know any better, you could say that Beeple's art looks like something AI could spit out after a few days of good honest work.
But there you'd be wrong: Winkelmann said himself that for his most popular pieces he was using "Cinema 4D and mostly Octane", which is a 3D graphics render software dating back, well, before AI was making waves in the art world. And just like with AI-generated pictures, there was a time when 3D rendered imagery, such as the one used for video games and advertisements, wasn't considered worthy of the 'art' label, too.
Speaking on the matter of AI's impact on the art world, Franco Reda, who co-founded The Creative Lab and is the person responsible for a handful of superhero munchkins you see on this list, believes people just need time to adjust. "I strongly believe that it’s enabling new forms of creativity and enhancing existing ones – there’s nothing to fear!" Reda told Bored Panda. "As with all emerging technologies, there are always concerns about the impact on human creativity and artistic expression."
Franco is right. If you're familiar with the history of art, you might remember that every technological breakthrough 99% of the time is greeted with public concern. When computers arrived in the early 1980s, so did a new kind of anxiety called 'Cyberphobia'. The same thing happened with photography, which took decades to be recognized as an art form. At one point, it was even called "the “mortal enemy” of art. Not only was it seen as something that wasn't quite an art, but also posed a serious threat to replace it. Sounds familiar?
But let's move on and talk about the act of creating. Something that both Franco and Milan are recognized for online. "It took about five or six iterations before I was completely happy with the result. The hardest part of the process was finding the correct prompt. Getting it wrong can often lead to some bizarre renders (and some happy accidents)," Franco said about his baby Indy, sounding like a real Bob Ross of tech art.
Meanwhile, for Milan, one of his AI-generated videos can take "anywhere from 4 hours to 2 days" to finish, with 50 to 100 runs of the exact same prompt for images. "It depends on how well AI works with me, as it is still unreliable," said Jaram, who's been mastering Midjourney for over a year now, ever since it was still in beta.
"Basically, I will run a few prompts over and over and if it doesn't jive with what I see in my brain, I will rewrite the prompt as detailed as I can. Sometimes it can take a few hours to nail it down. Some prompts are paragraphs long," he explains, saying that the current version of the software prefers artists "describing the scene as you would in a book."