No, Generative AI Is Not Art

Not saying it's wrong or bad, it's just not art. Come at me.

I had a different piece all lined up for this week but then I saw something annoying on social media that made me throw the draft on the backburner and settle into a nice cleansing rant. This works out fine anyway because the council has decided that the patch of road directly outside my house is due for a roadworks project to shame Rome at its height, and my peak writing hours are, for the time being, accompanied by the gentle chorus of what I can only describe as a chainsaw fight on a commercial tarmac.

So this will do nicely. Measured essays on current affairs require a level of concentration that I can’t fully dedicate myself to right now, but rants come out smooth and easy and are nothing if not helped by the sound of kaiju sized mechs resurfacing an obsidian plane three inches from my fucking window.

So let’s stir some controversy and get the blood flowing. Let’s talk about art.

I made the mistake of reacting to this meme earlier in the week. I suspect it was made to be tongue in cheek but the person posting it was dead serious.

As is standard for this meme format, the shorthand is that you know the correct opinion by the fact that it is espoused by the very handsome bearded man on the left, whereas the incorrect opinion is given to the ugly crying bespectacled character on the right who I am reasonably assured is named Soyjak.

Now, for context, the painting that Soyjak is valiantly trying to defend in the upper right is called Vir Heroicus Sublimis. It was created by the abstract expressionist Barnett Newman in 1951. Being that it’s over 70 years old and even the artist has been dead at least a decade before most of its living critics were born, it’s starting to stretch the definition of modern art. But then, modern art isn’t even mentioned here. In the words of the art cynic who brought this to my attention, this is just bad art. The only thing it reveals is that Barnett Newman was a lazy, pathetic failure.

These feathers are ruffled, of course, in defense of AI art, a defense laid down by newsletter authors whose catalogues are illustrated entirely by Midjourney images.

I can understand becoming frustrated if people razz you all the time for using these images. I’m not going to point my finger and shriek at you, I use them too. My Adobe package comes with Firefly, the idiot cousin to all the ones you’ve actually heard of, as well as like 57 other apps I’ll never use. I think header images are very important and I try to assemble them myself where I can, but sometimes I have a clear vision of something that just isn’t within my skill set, so yeah, I tell the robot butler to do it.

What it spits out is mostly garbage and never as good as the ones I’ve done myself. I’ll grant that I’ve made some pretty cool things using partial AI. This one’s quite good:

This article was about outdated social media platforms refusing to die, and I really wanted people ignoring a blue zombie bird—but I couldn’t Photoshop that and there weren’t any zombie birds in my stock library so I roboshopped the bird into a stock photo and touched it up.

This, on the other hand, was entirely generated and is what I’ve seen referred to online as “AI slop”:

I’ll admit my own follies, I think it’s a virtue, but I wasn’t going to miss my publishing deadline labouring over this psychedelic idiocy any longer. In retrospect I’m sure there was something else I could have used here. This piece was about Nazis online who use frog avatars, and I just had this image stuck in my head of these evil piece of shit frogs climbing out of the internet and infecting the world. The image in my head was much better than this.

Generative AI enthusiasts will tell me if I’d used a better prompt and a better software than the offensively overpriced PDF company then I could have pulled something out of my hat here, but no matter how much I cleaned this up and found a machine that knew how many legs a fucking frog has it would still have looked like what you see when you take a handful of shrooms and then chase it with high impact cranial trauma.

All AI images look like this, and no, they’re not art. Even when I’m using them, and using them well, like in the bird image above, the best I can say about them is that they are design.

---

Just interrupting to let you know the vast majority of what I publish is free, but if you wanna upgrade to a paid subscription for just $5 a month ($50 for a year—cheaper!!), not only do you help me continue doing what I love, but you get every article a whole week earlier than everyone else.

Don’t want to subscribe via Substack? A Ghost version is also available for paid subscriptions only.

---

Why aren’t they art? There’s no artist, for one thing. Who is the artist? I’m not the artist. I commissioned something, maybe. The computer damn sure isn’t the artist. Don’t make me explain to you why a computer isn’t an artist. It doesn’t have any intentionality or vision.

Look, I could put together a performance art piece that’s just me kicking a bowling ball down some stairs. There’s no doubt in my mind that the art cynic who posted the above meme would consider it to be really bad art, but categorically speaking if I set out to do this as art and I had something I was trying to express in doing it then there’s not really anything else you could call what I’m doing.

But if I trip and kick a bowling ball down some stairs by accident, then that’s not art, it’s just some shit that happened. If you describe to someone else seeing me kick the bowling ball, you’re not an artist. The ball isn’t an artist.

Generative AI isn’t art, it’s just some shit that happened.

As design, it’s fine. It’s great. What I’m doing with my header images is trying to attract people’s attention enough that they’ll hopefully engage with my real art, which is the writing. You don’t have to say this is good art either, but it’s unquestionably what this is.

But then there’s that accusation also, that Vir Heroicus Sublimis is bad art, or Jackson Pollock, or that banana someone duct taped to a wall. Duchamp. The Piss Christ. Picasso, for that matter. Is this bad art?

I hesitate to even present an image of Vir Heroicus Sublimis because its context and physicality and how you feel standing in front of it is actually part of the artwork. I might as well show you a still from a movie and ask you if it’s a good movie. You’re not going to experience it in a picture of it and showing it to you will only feed into the narrative that it’s the work of a lazy pathetic individual who got lucky dumping some red paint on a canvas and convincing people to hang it in a gallery instead of using it as a ping pong table.

The best I can do is show a picture of other people looking at it, to give you a sense of what it is.

Now to address the stupid meme again: Soyjack’s strawjack argument is that Handsome Chad “just doesn’t understand” this painting. The funny thing is, nobody who understands this painting would ever mock somebody for not understanding this painting. There is very little to understand. It’s an expressionist piece, basically all it is doing is transmitting an emotion from the artist to the observer.

The art cynic of course hears very little of this before breaking down into a series of scoffs and exaggerated wanking gestures. They look at this photo and all they can see is a big fat emperor striding through the gallery with his wang out. How can we possibly explain this object having any actual value to anybody in any scenario that doesn’t involve (1) Hipsters playing pretend (2) Money laundering, or (3) Brainwashing? I saw someone repeatedly insisting modern art is a CIA PSYOP.

After all, what is this, really? It’s not creative. It doesn’t show off a high degree of technical skill. You can argue it isn’t even aesthetically attractive. So what good is it? How can anybody possibly like or value this?

One day in the distant future I dream of a world where people have finally stopped talking about this fucking banana

The question “what is art?” gets bandied around a lot by all classes of people, whether they’re mocking art, artists trying to position themselves, or just genuine academic curiosity. For me (and I think this is close to the correct answer because I think everything I think is correct or else I wouldn’t think it—don’t lie, you are exactly the same): Human to human emotional communication is the most important thing about what makes something art.

Not informational communication. Plenty of things convey information that aren’t art. You can communicate both information and emotion and that can be art, but the emotion is both a sufficient and necessary condition. That’s why this meme makes me want to cut a bitch. Handsome Chad doesn’t understand Newman’s red panels, who cares? Nobody needs him to understand shit.

Bro, you’re standing inches from a stark, severe wall of deep crimson, not uniform, but visibly and at times violently applied by hand by an artist’s brush, you can see the strokes. It looms the fuck over you. Understand nothing. Feel.

Even if you’re only thinking of things in terms of pure biology and dismissing all the wanky artyfukingfarty, you have to admit that something like this will scientifically fire your neurons all over the damn place.

This was meant to communicate the emotion of pissing you off, and judging by the reaction it still gets from art cynics 107 years later, I think it’s relatively successful

That’s where generative AI completely stands by itself. Sure you can feel something when you look at something a computer spat out, but the feeling is incidental. I feel something when a passerby gets hit in the groin by the bowling ball I kicked down the stairs. The computer is not transmitting some feeling to you because it is not feeling anything.

“But the prompt!” I hear you scream. Surely that is where the artistry lies. So many generative AI enthusiasts will insist on the prompt, and I guess if I’m even slightly receptive to any argument then this would be it. But man… The prompt is basically you talking to the computer algorithm, so you have to be good at knowing how the algorithm works and… you’re still just talking about technical skill here, right?

I kind of suspect that people in the center of the venn diagram of being cynical about modern art, being nostalgic about classical art, and being enthusiastic about generative AI, are sort of thinking about all art in the way I think about the header images that I use for my newsletters—in short, what art comes down to is either good or bad design. What utility does it serve? How much technical skill went into producing it?

You can look at Vir Heroicus Sublimis and see only Barnett Newman’s pathetic failure as it clearly only exists because he lacked the technical proficiency of Renoir. But then when you talk about your skill with formulating a Midjourney prompt I’m kind of just thinking you’re getting above average results out of a clunky and imprecise English-based programming language. Not going to hold that against you, though.

Maybe we can all find common ground laughing together about how we’re debating the nature of art based on issues raised by a fucking Soyjack meme created in MS Paint in 50 seconds, tops.

Enjoyed reading this but don’t feel like committing to a subscription? A one time tip would be greatly appreciated!