Artificial Intelligence is Upending the Art World. Are you Next?

We are urgently asking ourselves a question never definitively answered: What is art?

Théâtre D’opéra Spatia by Jason M. Allen. Or a software program called Midjourney. I’ll let you decide.

I would like to begin this Cappuccino in the most metafictional way possible, which is by complaining that technology has ruined, or is actively in the process of ruining, everything.

No, the irony isn’t lost on me that I am using technology to convey this message.

I may have used hyperbole to launch my argument into the virtual broth of cyberspace, but that doesn’t mean I’m wrong. The problem with technology is that we’re in charge of it. Us. We’re like children who’ve just gleefully discovered that a magnifying glass incinerates ants.

One by one, we’re exterminating “democratizing” the arts by developing software that enables anyone to create, and on the surface of it, that’s a fine and noble thing. Creating art should be accessible to everyone. Where we seem to be running into trouble—quite a lot, in fact—is when the ease of production or reproduction squeezes “real” artists out of a job.

By “real,” I mean someone who has devoted many if not most of her waking hours to mastering her art. Through discipline, drive, and endless years of practice, of being very bad at something until she becomes very good, an artist pays her dues. As a result, she acquires the self-confidence of mastery and a strength of character borne of discipline and adversity, followed by what is hopefully an earned reward.

These are the unquantifiable benefits of discipline that are just as real as love (equally unquantifiable). Just because you can’t tote them up like a column of numbers doesn’t make them any less real.

Now, apparently, all you have to do to create visual art, mix music, write a novel, craft a movie, or sculpt a statue is press a button. All the work of imagination is done for you. No effort required.

Is that what the democratization of the arts should be? Letting a machine do all the work while you take all the credit?

Let me give you one of ten million for-instances why we’ve solved one problem (dispensing with the time it takes to master an art) and created three others.

In the not-too-distant past, recording an album, say, meant recruiting a variety of musicians to play your music. Ideally, these would be seasoned professionals who could read a chart (sight-read sheet music). Anyone who had achieved that level of musical proficiency would have obviously studied his instrument for years, and would now be paid to do what he had learned to do best, which is play music. Let’s say there were six musicians in total that you hired, plus the cost of a recording studio and sound engineer. That’s a sizable expenditure of money, but with economic benefit to many (as well as yourself).

To support the arts, but you must invest in artists. The math is simple.

Thanks to Progress™, that’s all unnecessary now. Even a low-end synthesizer can recreate the sounds you’re looking for with just a few key commands and the push of a button, eliminating the need for acoustic musicians. In fifteen minutes, you can lay down tracks that it used to take fifteen years of rigorous study to achieve.

That’s progress, right?

Artificial Intelligence is coming for the visual arts, too. In a recent New York Times article, “An A.I.-Generated Picture Won an Art Prize. Artists Aren’t Happy,” it was revealed that a man named Jason M. Allen of Pueblo West, Colorado, won an annual art competition with his visually arresting piece, Théâtre D’opéra Spatial.

The trouble was, he didn’t paint it.

Théâtre D’opéra Spatial is the result of entering a handful of keywords (Allen won’t disclose what those words were, but “space” and “opera” are two good guesses) into an artificial intelligence program called Midjourney, which turns lines of text into stunningly realistic graphics. I tried it myself, the results of which I will share with you later on in this article.

Allen didn’t lie about the work’s origins. He submitted it under the name “Jason M. Allen via Midjourney,” which exonerates him (in my opinion) of any wrongdoing. More culpable—again, in my opinion—is the Colorado State Fair’s annual art competition committee for not making the distinction between digital art and art that’s created by typing in a few random keywords. Tsk tsk.

The general outcry against giving the ribbon to a guy who typed in a few keywords to create his masterwork was immense. “This sucks for the exact same reason we don't let robots participate in the Olympics,” tweeted a responder named Fluxophile. From another tweeter who goes by the handle 中山雅紀＠僧侶･数理美術家, “The value of art exists in the ‘PROCESS,’ not the ‘RESULT.’ Art has no value if the artist cannot explain the ‘PROCESS’ of how it was created. Thus, no matter how novel the AI’s drawings may seem, they are not art.”

These are legitimate ethical questions being raised. Midjourney and its fellow image generator, DALL-E 2, operate by scraping millions of images off the open web, everything (presumably) from Rembrandt to Jeff Koons. The algorithm is programmed to recognize patterns and colors so it can generate images on command in a style similar to what it’s learned. The results are astonishing.

Because I didn’t feel it was ethical to hold forth on Midjourney until I tried it, I signed up and gave it a whirl. The interface is simple enough, although I wouldn’t call it intuitive. In my first go-round, I typed in the words “Tokyo in year 3031” and within less than a minute, Midjourney gave me this:

Interesting. A little bit Jetsons meets Graveyard of the Fireflies. So I tried something a bit more psychologically complicated. This time, I typed “Carnival of lunatics.” Here’s how Midjourney interpreted those words:

“Lunatic” is a strong pejorative, which is why I was curious to see what Midjourney might do with it. Clearly, the image in the upper righthand corner best exemplifies the spirit of lunatic, although to be fair, all these carnival people look a little frightening to me. We’re seeing them through a sickly vapor. We’re either imagining this freakish demimonde or we’ve drunk too much absinthe.

RJ Palmer, a digital artist who you might assume would enjoy Midjourney and its brethren has some interesting things to say: “The results can be surprising and funny and beautiful, but only because of the vast trove of human creativity it was trained on … What makes this AI different is that it’s explicitly trained on current working artists. This thing wants our jobs, its actively anti-artist … It opens profound questions about the ethics of laundering human creativity.”

There are the usual dissenters who are rightly concerned about the manufacture of sexual and/or violent content, especially that involving trademarked figures and celebrities. Frankly, as far as copyright goes, now that Google is here and the genie is out of the bottle, I don’t think she’s ever going in again. The Innerwebs are too vast, too unregulated, to control things like intellectual property. Midjourney and others are most assuredly pillaging the work of others, work for which the artist(s) will never be attributed or compensated. This, now, is our Brave New World.

Do I think Midjourney and other artificial intelligence image generators are going to put traditional artists out of business the same way samplers and synthesizers have eliminated the need for real musicians? At this juncture, I doubt it. Creating “art” by typing in a few keywords is fun, but it didn’t feel as though I was actually making anything noteworthy. It was more like a game. Since anyone can do it, there is no particular exclusivity or status attached to it, which means no attendant bragging rights. After all, what is there to brag about—that you concocted a series of clever keywords?

Having said that, I don’t believe artists are overreacting. Artificial intelligence is already driving a lot of creators out of business.

AI can write novels, so there’s some more “good” news. No more pesky writers failing to meet their deadlines or getting sick or demanding royalties. Pretty soon, any respectable publishing house will be able to churn out as many titles a day as it wants. Think I’m joking? I’m not.

Oh, and chances are, AI is coming for your job, too. The World Economic Forum estimates that 85 million jobs will be replaced by machines with AI by 2025.

Lemme know how that pans out for you.

The wholly unapologetic Jason M. Allen of Théâtre D’opéra Spatial fame had some chilling words for his detractors. “This isn’t going to stop,” he told The New York Times. “Art is dead, dude. It’s over. A.I. won. Humans lost.”

As much as I’d like to believe that isn’t true … there are, unfortunately, zero reasons to think otherwise. Humankind has an unblemished track record of hubris, arrogance, short-sightedness, and appalling disregard for the better wisdom of “just because we can doesn’t mean we should.”

Of course we’re going to automate art. We already have. What remains to be seen is what kind of war artists will mount against the rising tide of AI.

I really hope we win.

Copyright © 2022 Stacey Eskelin

I’m quite curious to know what you have to say, so be sure to leave your thoughts in the comments section below.