Let's Talk: AI Slop

I'm not going to sugarcoat it: I hate and despise AI art from the literal depths of my heart. Every myofibril, every sarcomere, every individual actin and myosin filament, are drenched in my utter disdain for it. It makes me want to projectile vomit every ounce of food, drink, formula, and placental fluid that I have ever consumed throughout my existence. It carries a moral stench that has no equal. Imagine an entire colony of cats urinating on the same patch of carpet for five years straight. Next, imagine that this patch of carpet is never cleaned during this entire period. That odor would still be nothing compared to the absolute blasphemy that is AI art.

However, I am not writing this piece solely to rant. I've made it abundantly clear that I hate AI art. Now I want to tell you why I hate it and why anyone reading this should also hate (or at least oppose) it and take action to reduce and eliminate its use. I've blown off my steam and so from here on out, I will be speaking with a clear head.

I realize that not everyone shares my passion for the arts. Because of that, I will start with a more universal concern: the environment. Generative AI uses immense quantities of water. There are three main factors creating these demands: Data centers that need water to cool their servers, thermal power plants that need water to cool their systems (electricity is used to cool data center servers in water-scarce areas), and chip manufacturers that need water for the manufacturing process.

According to research done by The University of California and The Washington Post, it takes 519 milliliters (2 cups) of water for Chat GPT-4 to create a single, 100-word email. Generating one of these emails each week for a year uses 27 liters (7 gallons) of water. And if 1 out of every 10 working Americans were to do this for a year, that would require 435,235,476 liters (114,977,049 gallons) of water. That is equal to the amount of water used by all households in Rhode Island over the course of 1.5 days.

Coming from an electricity perspective, the same 100-word email consumes 0.14 kilowatt-hours of electricity, or the amount of electricity needed to power 14 LED lightbulbs for an hour. One email weekly for a year requires 9.5 kilowatt-hours. This could instead power 9.3 Washington D.C. households for an hour. Done by 1 out of 10 working Americans, this would use 121,517 megawatt-hours, or the amount of electricity used by all D.C. houses for 20 days. For areas with AI data centers, the cost of this electricity use is of course pushed onto local residents.

And that is all for a 100-word email that someone could have just banged out in a minute. For more substantial AI queries, such as essays or artworks, the resource cost is even more immense. I would strongly encourage any readers to check out the original Washington Post article, as it contains some additional information as well as some great visuals.

I would be remiss not to discuss the uses of AI art that are not so much actively malicious as they are uniformed. These instances aren't game developers looking to cut corners or gallery curators shutting out real artists from spaces. These are everyday folks using AI for a quick laugh. They would otherwise not be "creating" anything and get about five seconds of amusement from their AI work before tossing it aside.  They feel no level of investment in their quick work and do not have a desire to see it properly brought to life through commissioning someone. I think the most popular example of this as of late would be the AI action figure and accessories. 

If AI art in the advertisement, entertainment, and business industries is a metastatic cancer, then AI art used by laypeople is precancerous. The large intestine of everyday life is covered in these precancerous, AI art polyps. And sure, a polyp or two is generally not terribly worrisome. They're a pretty common colonoscopy finding in older folks. But when someone's colon is absolutely bursting at the seams with polyps, that's actually classified as a disease called Familial Adenomatous Polyposis. Each individual polyp has a pretty low risk of becoming malignant, but when you've got an absurdly massive number of them, the risk becomes pretty huge that at least one of them becomes cancerous. And the assumption in medicine is that if a person with this disease does not do anything about all those polyps, they WILL develop colon cancer. 

AI art's use by laypeople is the same situation.  A single person's action figure post is not doing a significant amount of harm to the art world. But with absolutely everyone making these works and people scrolling through AI art post after post, AI art is becoming eerily normalized and the general populace desensitized. If this trend continues, what is considered generally unacceptable today (for example, generating an impressionist-style painting and presenting it as if you actually painted it) may unfortunately become acceptable in a year. 

Society has already begun falling quite far down this slippery slope, as the Oscars committee recently announced that going forward, AI generated films would be eligible to win said awards. If the members of the Oscars committee, who are deeply entrenched in the world of film and art, have already been swayed to such a degree, then the general populace who is not deeply invested in the world of art does not stand a snowball's chance in hell. Education on the problematic nature of AI art is the only way to improve its chances. Unfortunately, this is just not going to happen. Most people just do not care about the arts enough. I work in healthcare and it is an unbelievable struggle to get the average person to care about their wellbeing and to care about keeping themselves from dying of preventable illnesses. So getting them passionate about something that is without a doubt less important than their continued existence is, in my opinion, like trying to light a fire underwater. And due to this, there is no "safe" level of AI art in general society. There is no parts-per-billion that will not multiply into a toxic tsunami.

AI art is also creatively bankrupt and plagiaristic by nature. The nature by which AI programs compile art makes crediting the original artists impossible. Furthermore, virtually all AI compiling is done with uninformed consent or even no consent at all.

There is a fairly popular and humorous misquote of Japanese animator Hayao Miyazaki claiming that "Anime was a mistake, it's nothing but trash". Most people familiar with the quote are aware that it is purposefully inaccurate and used as a joke. However, they never tend to talk about the content of the actual quote, which is as follows:

"You see, whether you can draw like this or not, being able to think up this kind of design, it depends on whether or not you can say to yourself, "Oh, yeah, girls like this exist in real life.”

If you don’t spend time watching real people, you can’t do this, because you’ve never seen it.”

Some people spend their lives interested only in themselves.”

“Almost all Japanese animation is produced with hardly any basis taken from observing real people, you know. It’s produced by humans who can’t stand looking at other humans. And that's why the industry is full of otaku!"

Miyazaki's sentiment translates remarkably well to the issue of AI art. AI art takes no inspiration from real people, from real life, from real situations, places, and experiences. AI programs are trained using preexisting art pieces. And once those programs spit out pieces, said pieces are used to further train AI programs. Throughout the cascade of AI art generation, there is no point at which a new, fresh, personal interpretation of the human experience is injected into said piece. AI programs cannot be inspired. They cannot have muses. They cannot reflect on experiences, love, fear, hatred, or fascination. They can only copy, repeat, and regurgitate.

But these varying interpretations are the genes and alleles of art! Without them, AI art is pulling from an exponentially shrinking gene pool and is practically inbred. Humans may have depicted the devil millions of times in art, but no single depiction has been the same as another. A tyrant who revels in suffering and punishment, a somber and misunderstood figure, a really sexy demon- there is no end to variety of ways in which humanity has interpreted, and will continue, to interpret the devil. And that's not to mention the choice of medium, texture, color, composition, pattern, value and the rest of the sea of stylistic choices that art offers. The human experience and the ways in which we can express it are boundless. AI art, on the other hand, is bound and confined within a cage of parameters, algorithms, and plagiarism.

Of course though, if I'm talking about plagiarism then I need to acknowledge the elephant in the room. Human artists are not immune to copying. Actually, almost every single person on this planet has likely copied art. Did you grow up copying poses of Pokemon from a guide book or characters from a comic book? Did your parents buy you a "How to Draw Disney Characters" book with step-by-step instructions? Did you get bored in class one day and decide to copy the dinosaur design on your social studies folder? If you answered yes to any of these questions, then guess what? You have copied someone's art!

But those instances are totally benign, as are many other situations. Kids copying poses just want to draw what they like. They also aren't publishing their works or making money off them. Actually, nobody really cares at all if someone copies their art as long as the copier doesn't take credit for the concept or composition, acknowledges and credits the original artist, and doesn't try to monetize their derivative work. Copying only becomes plagiarism when someone doesn't acknowledge the original artist and tries to pass the work off as their own idea. Copying in and of itself is an excellent way of learning and should never be shamed or discouraged. After all, making master studies is a valid and even encouraged way for serious artists to hone their skills; attempting to perfectly replicate the composition, brush strokes, lighting, and colors of a famous work can help fine tune one's skills.

If I had to guess, I would say that most contemporary artists got their start with an assortment of "How to Draw" books.

If someone does copy a work and attempt to claim the composition as their own or monetize it though, then that is absolutely reprehensible. These situations are unethical uses of copying and enter the realm of plagiarism. They should not happen, they should not exist, and shame on anyone who creates them. I will not hesitate to call out anyone for plagiarism, and neither should anyone else. Yes, AI art is vile, but so is human plagiarism. Hating them is not mutually exclusive.

Besides straining resources and stealing work, AI art is of course also eliminating jobs for artists. Game and animation studios, book publishers, large companies, they are all using AI generated artwork to avoid paying actual artists to work for them. It's repugnant at every level, but especially so when it's done by small businesses. These businesses depend on the good will of locals who acknowledge that although they will likely pay more for a product than they would at a large retail chain, they receive a product and service with fifty times the love, service, and passion that a chain could ever hope to provide.

How hypocritical is it then for these same businesses to ignore the passion and talent of local artists and instead generate AI art? I was particularly frustrated last fall when I saw that a local haunted hayride opened up a daytime ride for small children but opted to use AI to create the advertising. A haunted hayride is entrenched as deeply in the arts and human collaboration as it gets. It depends on textiles and costume-making, SFX, makeup, props, acting, dance, music- almost every form of art. Its directors are fully aware of how dependent on art and humans their production is, yet they couldn't muster the respect to hire a graphic designer for their new attraction.

I have similar feelings of disgust toward Michaels, who was found to be selling AI generated art in its 2024 Halloween collection. An arts and crafts store of all places should be supporting artists. They could have very easily commissioned a few artists to supply them with designs for their collection. And because the majority of artists are pretty strapped for cash, they could have likely done it for a pretty cheap rate. The bad publicity and sales they lost due to this scandal definitely cost them more than what it would have cost them to commission said artists.

Granted, a spokesperson from the company stated that they purchased the art from a vendor who licensed the original source material from an artist. Prior to selling the design to Michaels, the vendor added an AI layer to the image. This representative claimed that Michaels was unaware of this addition and that it was an "unacceptable error". They then doubled down on Michael's commitment to human artists. However, these words turned out to mean nothing. On Michael's online craft-selling platform MakerPlace, you can currently pay $100 to enroll in a class about making and selling AI generated products. MakerPlace also offers a tutorial on how to turn AI generated art into bookmarks.

I've also seen many people defending AI art distort the concept of accessibility. AI proponents attempt to defend AI art by saying that it makes art more financially and physically accessible to the average person. However, this defense confuses accessibility with ease. Art in its most very basic form is accessible to someone from any walk of life.

Let's start off with finances. Art is not just big, fancy canvases brushed with even more expensive oil paints or sculptures made from premium clays and carved with high-end tools. It's also doodles scratched onto scrap paper or homework, shapes drawn in the sand or dirt, and fallen leaves arranged into different patterns. Now that's not to say that art never has financial barriers- as previously mentioned, oil paints can get very expensive! So can lots of other supplies. There's no arguing that very high-quality art tools and mediums are, for the general population, unaffordable.

However, in many cases, the tools don't make the artist. The artist makes the tools. If you hand a novice a set of 48 Schmincke Horadam watercolors (MSRP $850) and a master a set of 16 Crayola watercolors (MSRP $4), we all know who will produce the nicer work. Very fancy tools and mediums can absolutely elevate the work of a skilled professional who knows how to take advantage of their unique characteristics. But for the vast majority of the population, which includes the most vehement defenders of AI art, such people are nowhere near skilled enough or knowledgeable enough about art theory to make proper use of expensive tools. They are just as well off using lower quality but affordable tools.

Just take a look at Kirsty Partridge using Crayola colored pencils to make an impressive portrait of an owl. These are the exact same colored pencils used by kindergarteners to draw stick figures and suns in the corners of pages. The difference is in the artist!

Regarding physical disabilities, this is not an issue either. Artists with all sorts of physical disabilities - blindness, malformed limbs, a complete lack of limbs - have found unique ways to create art. Take artist Noah Matthew Howard for example. He was born without any arms and grew up learning to use his feet to draw, paint, and write. Over time he has developed an incredible level of dexterity with his feet that allows him to create fine and precise details such as the eye of a moose or the lights on a carousel. Noticeably though, while Nolan lacks arms, he has an abundance of something that is completely lacking in AI artists: motivation and work ethic. 

He is drawing with his toes. You have no excuses.

I 've Been Tricked Again, 2021

Crowns of Courage, 2021

It's quite simple: AI artists don't want to learn how to make art. They want to create art but don't want to learn shape, form, proportions, pattern, balance, value, or any of the principles of art. They don't want to learn color theory, or how to gesso a canvas, or how to blend colored pencils, or how to bake clay without burning it. Imagine if someone told you they wanted to join a local basketball team and become a great basketball player but didn't want to learn how to dribble or shoot the ball. Or if someone told you they wanted to repair classic cars but didn't want to learn how different tools work or what each part of the car does. That would be absurd! And AI art is no different. Everyone has the means to create art, but AI proponents don't want to get good at it. That's not an accessibility issue. That's a motivation and work ethic issue. 

But you know what's absolutely crazy? Products for people who don't want to dive down the entire rabbit hole of art already exist!  We have coloring books, paint-by-numbers, pre-sculpted, ready-to-paint ceramic model, and similar products. 

The difference is though, folks who use these products recognize that they are taking significantly abbreviated artistic journeys. They acknowledge that they didn't have to worry about composition or line weight. They know that a lot of that work was already done for them. And when they finish their work, they don't present them or submit them to galleries as if they are completely, 100% made by them. They take pride in the part of the process that they did play a role in, but they don't try to pass off a paint-by-number's or coloring book's composition as their own. They don't try to claim that they sculpted the flower pot that they just painted. They don't take credit for work that they didn't do. 

Who would have thought that a product as inexpensive and widely-available as coloring books could solve this whole "I don't want to get good at drawing" dilemma?

And that is totally fine! I am all for abbreviated art journeys. Many people just want to relax and unwind a little. Most people aren't mentally unstable artists who have a voice in the back of their head telling them to make something every day and make sure it is better than the previous day's. They are casual enjoyers of art and they should absolutely continue to have opportunities to stroll through the world of art at their leisure.

The problem with AI artists, however, is that their art journey was not even abbreviated. It was just flat out non-existent. And despite this journey being non-existent, they present their work as if they were responsible for every facet of their work's existence. It really just boils down to the fact that you shouldn't claim credit for something you didn't do 

Outside of the visuals of AI art themselves, one argument I hear from some is that creating the prompts for AI art is a skill that needs to be grown and refined. I don't entirely disagree with this. Many pieces of art start not with sketches but with written descriptions. Many of my pieces start as blurbs in my phone's notes app. I describe the scene, the characters, the mood, what's going on, the color scheme, etc. Having a little written description provides me with some helpful direction when I go to sketch a piece. Writing these descriptions is absolutely a skill totally separate from skills in drawing, painting, etc. 

But that's the thing: it's a separate skill, a writing skill. And writing is not the same thing as visual arts.  Yes, by making AI art someone will develop some abilities for writing descriptions of scenes or projects. But they aren't developing any skills in visual art. And ultimately, any written prompts are only a very small part of the creative process. Someone partaking in 2% of the process does not give them the right to claim that they actually partook in the other 98% as well.

The last point I want to share is a bit more emotional than logistical or ethical. Visual arts are usually not considered performative (please note the emphasis on usually, as there are absolutely some performance-based works of visual art). For most pieces though, audiences are primarily concerned with the end product. But visual arts still have a process. Every stroke you make and every color you choose are like pirouettes and promenades in a dance routine.

The artistic process is filled with joy, sadness, fear, frustration, and awe. Someone can experience a rollercoaster of emotions throughout the process of creating a piece. They might feel anger that a certain composition they imagined doesn't look good on paper. Later on they might feel overjoyed when a color they didn't initially consider actually looks good in their painting. Every piece I make has stories like these behind it. I can tell you when I felt fearful, when I felt relieved, when I had a "Eureka!" moment, when I wanted to rip up my work and burn it. For some of my works, the story behind them might be more interesting than even the work itself. 

But with AI art, there is no story. Just a blip of an ending. The person who generates AI art has nothing to look back on or reflect on. There is no plot, no conflict, no rising action, no climax. To skip out on this experience is to summit Mt. Everest without climbing it or to cross the finish line for a marathon without running it. It is borderline antithetical to the idea of art. Art doesn't just start existing, it is created. It is nourished, cultivated, and tended to. Yet AI art aims to eliminate the act of creation as if you could eliminate the nine months that a fetus spends in its mother's womb.

-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

You can view Noah Matthew Howard's art here: Armless Art

You can view Kirsty Partridge's art here: Kirsty Partridge – Helping You Make Art You Love!