The Light at the End of the Tunnel

Implications and impact of generative AI in the visual arts

Photo by Claudia Soraya on Unsplash

Over the past months, I have thought a lot about the raging generative AI art debate. When AI art apps like Wombo, Dall-E, and Midjourney first burst on the scene, my reaction was one of visceral unease, coming from a place of bewilderment and fear. I agreed with many that this powerful technology can potentially replace artists and further degrade our abysmal culture. Now that I’ve had more time to consider the impact, I’m finally ready to share my observations.

Those of you who have been holding your breath in anticipation may now exhale in relief.

It’s hardly groundbreaking to observe that every new development in technology, especially one as disruptive as AI promises to be, is greeted with fear, resistance, and dire predictions from some portion of society. The other portion (usually the majority) greets it with open arms and gobbles up the attendant pie-in-the-sky promises made by developers. Both responses are short-sighted and simplistic; I wish to explore some nuances here. Warning: this is a subject that brings out strong emotions, so if you are already getting irritated, please bear with me and read to the end; you may be surprised. As always, I welcome respectful comments.

To begin with, as others have already pointed out, the debate over the worth of AI art won’t get far if terms aren’t defined and agreed upon by the opposing parties. To this end, I humbly offer my definition:

Art is the nebulous and inadequate term for mankind’s attempt to articulate the Real through whatever tools and medium are available. The Real can be something small, unimportant, and transitory, or it can be as vast as the attempt to explain all of Reality itself. The tools, whatever they may be, serve the mission and are not an end in themselves. The goal and intent are what matter. A kernel of the Real (Truth)lies at the heart of even the most lowly advertising art. An advertisement for chewing gum and a High Requiem Mass lie at opposite ends of a continuum, a spectrum where everything is a form of art, but these creations do not share equal value or merit, nor do they deserve equal attention or regard.

So why does one product of human imagination and craftsmanship have more value than another? To determine that, creative works should be judged on subject, intent, and established standards of execution—criteria that have no direct relation to the tools used to create the works.

Considering the results and not the tools, there is no reason that connections between the mind of the instigator/visionary (the “artist”) and the mind of the beholder can’t be established through AI images just as well as other mediums, assuming that similar thought and care (not physical effort) are put into a given image’s composition and execution.

On a molecular or subatomic level, paint and light are not much different than the electrons moving through a computer processor and manifesting through pixels on a screen. Minerals used to make paint and minerals used to make silicon chips are not that different chemically. Photons striking pigment on a canvas, electrons in the computer, and electrical signals translating an artist’s will via thought and muscle movement are all cousins. They are part of the whole created universe that man uses to express himself and explore his place in that same Creation.

I think that according to my definition, AI-generated art is a form of art (or has the potential to become so) but perhaps only with a small “a”. Could it ever rise to the level of capital “A” Art? How important are the artist’s intention and the audience’s reaction in this regard?

That’s a much harder question to answer.

Is it possible for AI art to promote virtues such as beauty, truth, and goodness? Theoretically, it can, depending on the subject matter and how it incorporates established laws of composition, form, etc. An AI program can easily produce images based on traditional guidelines, images that can uplift the viewer. On X (formerly Twitter), I follow several artists who are generating some very interesting work, and I appreciate the time and effort they have put into these. When judged on imagination, creativity, composition, color, and technique, some strike me as coming close to genuine Art.

But I still wonder: Would any virtue in such an image be negated because it was created with AI? Does the intent of the artist (or “prompt wrangler”, if you prefer) outweigh the method of execution? Since we’re exploring the “transactional” nature of the relationship between the instigator and the viewer, how important is the viewer’s reaction at the other end of the process?

Let’s compare and contrast two images of the Blessed Virgin Mary.

Image 1: generated by the author on Midjourney.

The first image was one I generated in the AI program Midjourney. Unless I were to print it out on paper or canvas, it only exists in the digital realm- and now in the mind of the beholder.

Image 2: Mosaic by Marko Rupnik, screenshotted from the internet.

The second image is a photo of a real-life mosaic, formed of stone set into grout, and it’s designed and rendered following the conventions of Christian iconography.

When I generated my image on Midjourney, my deliberate, conscious intent was to make a picture that would uplift the viewer’s thoughts and help them meditate on Mary as the Queen of Heaven as she contemplates her Son.

Does it achieve that? I honestly don’t know.

Is it inferior because it is divorced from the kinetic, physical experience of creating a physical work of art? Isn’t there something vital in the actual process, in the smell of the oils, the swishing of brushes in their rinse water? The satisfaction of guiding black ink over white paper?

Yes, those experiences are certainly valuable, but they are also separate from the finished product. When I read a novel, I don’t care much what brand of computer it was written on or how many cups of coffee the author guzzled as they agonized over the second plotline. I only care about the end product, about getting lost in the story. As a fellow creator, I might find those incidentals interesting later on, but when I first read a work or look at a visual image, my immediate response is: what does this mean and how does it move me?

I think most recipients, most passive viewers of any kind of finished work, are more interested in the emotional response it provokes—their ‘feelings’— than in the artist’s experience or methods. Whether this is a good or bad state of affairs is the subject for another discussion.

And yet—

And yet, it’s undeniable that there is always a connection between creator and audience, however faint or unintentional, on some quantum or spiritual level. The commercial artist’s aim is to get the observer to open their wallet and buy the client’s product, and every stroke of the pen or brush furthers that goal. The propagandist’s aim is to uphold the current regime’s policies. Stained glass windows and painted plaster impart scriptural lessons via every glance that falls on them. Thus, a genuine connection is established between creator and recipient, no matter how nefarious, mundane, or lofty the aim.

To return to our Marian images: perhaps it’s not the use of a limited array of approved materials and techniques, tools, or effort that give an image its inherent worth. What are some other contributing factors? Is image No. 1 worthy to hang in a church, or at least in a Catholic home? Does my intent to create a reverent image outweigh the fact I used ones and zeros to generate it, instead of paint and panel?

Regarding Image 2: The artist, former priest Marko Rupnik, has been credibly accused of using his status as an acclaimed artist and cleric to manipulate and sexually exploit the vulnerable members of a female religious order. There is even evidence that he deliberately included elements of bizarre, blasphemous rituals in the creation of some of his works, many of which adorn churches throughout the world. (For a more personal account of how Rupnik’s art resonates with victims of abuse, please read Sr. Danielle Victoria’s recent post here:)

Does the fact that ex-Fr. Rupnik used his living, if sinful, human hands to place real stone in real grout mitigate the fact he was cynically manipulating the sisters under his influence as he created these pieces? While I’m not in the camp that holds “all artists must be pure or we can’t enjoy their work”, Rupnik’s loathsome behavior went far beyond the peccadillos of a typical brawling yet penitent Michelangelo. Going forward, could anyone who beholds Rupnik’s artwork forget about the sacrilege that attended its creation and truly meditate on higher things?

I know I couldn’t.

We can agree that the end result—an image that communicates something to the viewer, however produced—exists in both these examples; in that respect, they are both “art”. We now ask ourselves, setting aside the spiritual state and intentions of the minds initiating the images, which work is objectively better? That’s harder to answer, as we now come to a matter of education and personal taste.

I would contend that much of what we consider Art (however it’s produced) isn’t very good because it appeals to the lowest common denominator, much like the art of someone like Thomas Kinkade. While Kinkade engaged in shrewd marketing tactics, many folks genuinely love his soft-focus images of cozy cottages nestled amidst colorful landscaping. They sell a comforting ideal that demands nothing from the viewer. They are “pretty” and “realistic” compared to so much blatant ugliness produced in recent decades.

Although Kinkade’s work was created with real paint on real canvas, put there by a real hand, it’s not great Art with a capital “A”. It’s visual fast food, and we all know what access to vast amounts of empty calories has done to the American waistline, heart, and blood sugar levels.

Indeed, there’s already a sense that AI art is becoming a sort of cotton-candy-esque background to our daily visual life, as it pops up in memes and advertising art nearly everywhere one looks. It’s rapidly being accepted, but at the same time, it’s being ridiculed and tuned out as the novelty wears off.

And even with the better-quality, more thoughtful, “good” AI art (still small “a” for now), the more there is of it, no matter how good it may become, there is the real risk of people eventually tuning that out as well. When the market is flooded, the novelty wears off, and when a commodity becomes cheap and readily available, it’s no longer valued. Perhaps when that happens, the value of old-fashioned, hand-created work will rise.

Now let’s consider one of the biggest objections to the technology, the oft-repeated charge that AI trains on art “stolen” from real artists. This seems to be a gross oversimplification of how the process works.

The mind of a human artist is continually and subconsciously taking in information from the real world in order to synthesize it into finished works, and no artist would make any progress or develop their skill if they weren’t also studying the styles of other creators. Artists have always sketched in museums, copied from books, and studied other styles. Copying or being inspired by a style is not the same thing as copying a specific work of art stroke for stroke and passing it off as your own, which is rank plagiarism and theft.

Furthermore, stating that an AI is soulless and doesn’t know what it’s doing, while true, doesn’t really strengthen these objections. It’s still a form of rote training, whether that process is happening in the program itself or whether a programmer is scooping up reference material from the internet. Anyone putting their artwork on the internet has always known there is a very real chance it will be copied, at least in part, by other living artists.

Despite some artists’ claims, I don’t know if there has yet been a proven case where an AI program generated an accurate, complete, previously existing artwork and passed it off as original. (If any reader has hard evidence to the contrary, please—respectfully— chime in in the comments). To date, I’ve only seen artists upset that images are being generated “in the style of” certain living creators and that the AI was trained on works without permission or compensation.

News flash: I myself have “trained” on quite a few works of living artists. I have numerous Pinterest boards devoted to them. In fact, I recently purchased three volumes of a manga series because I love the artist’s style and hope studying it will infuse my own work with new and exciting qualities. That doesn’t mean for one moment I will reproduce his panels line for line. I’m certain those complaining have drawn much inspiration from their colleagues’ works, too. Copying styles is something every artist does as they try to figure out their own unique voice.

All this may be aggravating and frightening, but I’m not convinced it’s unethical or immoral. However, the actual legality may be a different issue. As a fellow author (I believe it was L. Jagi Lamplighter) observed some months ago, the current political climate might have a bearing on how juries view any upcoming cases. There is the possibility that the population’s inability to understand theories of machine learning, the ease with which they are swayed by emotional arguments, etc., may contribute to some sort of legal and cultural backlash against the technology.

The very emotion-charged question of whether or not this technology will lead to a flood of unemployed artists is complex. Of course, it’s a no-brainer that tons of people will use these tools to generate decent-looking, inexpensive book covers, images for marketing materials like short trailers, quicky websites, and so forth. So yes, some artists will lose work. But their case is not helped when advocates for hand-drawn/painted covers end up defending artwork that is, to put it charitably, not very attractive to potential buyers. Book covers are supposed to sell your book, not repel possible customers with strong amateur vibes.

If the final goal is to maximize profits, then yes, people will use these tools to help achieve those profits in the world of commercial art. It’s about scalability and economics. People will always complain that this is a side effect of capitalism, but capitalism operates like a law of nature—according to rather strict laws of supply and demand. People always want to pay as little as possible for what they need. Not everyone can afford to say, “I’ll have nothing but the finest hand-painted covers for my work,” any more than most families can afford the luxury of eating only grass-fed beef or wearing hand-sewn calfskin boots. In today’s world, the population relies on current methods of industrialized farming to survive, or people would starve to death. That may change as technology advances (and I hope it does), but it’s the reality on the ground for now. The same principle applies to mass-produced clothing, cars, and now, art.

However, I think visual arts should be held to a higher standard. What we’ve seen so far is that, along with scalability and mass-market production, comes a decline in quality. As with the flood of over-rendered, hyper-realistic 3D animation to which audiences have been subjected over recent years—the “Pixar Effect”— a backlash could include a return to more traditional, hand-drawn art styles.

Fortunately, there are already signs this is happening. Consider the popularity of the movie Klaus, which, although the creators used computer tools, had a traditional 2D look. Other recent animated releases (usually from independent studios) are returning to a more intimate, hand-crafted “feel” for their art styles. Thus, there is plenty of hope that there will always be at least some demand for traditionally produced art, but it might be a niche. With that in mind, it’s not too early for lesser-known or aspiring artists to figure out how to stand out against the colossal incoming wave of mass-produced imagery.

Anecdotally, I know of young people whose enjoyment in experimenting with AI inspired them to learn to draw traditionally with physical tools. I don’t think this is an isolated case. Many people might enjoy generating images with AI but will realize they lack the satisfaction of the physical process of creating art themselves. (Hopefully, this trend will solidify before Elon Musk’s Neurolink implants hit the mass market and we all evolve into flaccid, bloated brains, only capable of communicating via tech implants while our appendages wither into useless stumps).

Clearly, one the biggest dangers of AI is that it separates the creator from the physical process, from the exercise of the hand, the eye, and the body. We are not God, that we can speak being-ness out of nothingness with a word or thought. Our bodies need to be a part of the creative process. Making mistakes and correcting them within the constraints of time is good for us as creators. AI art lacks that tactile experience. Sure, you can fiddle with prompts and be disappointed (or horrified, or amused) and try new permutations of phrases until you get a result that matches your original inner vision, but it’s hardly the same thing.

That being said, I greatly dislike the mob mentality and tribalism around this issue—painting opposing sides as villains is not helpful. Ridiculing artists who have genuine fears for their dreams and their livelihood is cruel and shameful. Conversely, deriding all AI-generated images as “slop” looks like silly sour grapes when more and more images are genuinely unique and beautiful. Claiming the technology is “demonic” also does nothing for fruitful, reasoned discussions on its ethical use and impact.

I’m not convinced that generative AI tools are inherently less ethical or somehow outright evil when compared to traditional methods of creation. Pornography, blasphemy, lies, and propaganda have always been successfully produced, even with ancient tools. People are just as susceptible to the temptation to misuse their imagination and skills in pursuit of money or power as they ever were. Today, the only differences are ease of access and scalability. Is that bad? Perhaps, but that’s the reality we are now navigating.

At this moment in time, I don’t know what AI’s long-term impact on the visual arts will be. I don’t know how courts will rule, what the public will accept or reject, or which trends will rise and fall. Some artists will doubtless adopt an insular, Amish or Benedict Option-style mindset and double down on what is familiar to them. Of course, they are well within their rights to do so. Talented artists in this camp will end up riding the storm and emerging with more business than they can handle among a certain clientele.

However, on a practical level, I expect many other creators will decide for themselves how to incorporate it into their work and will regard it as one tool among many in ways that we can’t yet predict.

The light coming at us may be a new age of innovation, convenience, and beauty. Or it could be an unstoppable train of disappointment and mediocrity.

Whatever is coming, it’s too late to escape it entirely.

Author’s note:

As I was putting the finishing touches on this essay, the Vatican released a letter, Antiqua et Novi, addressing the possibilities and dangers regarding AI in many different situations and applications. I have only skimmed it thus far, but it seems to support some of my own observations regarding human creativity. It offers other valuable, well-reasoned, detailed insights as well, and I urge readers to take a look:

Antiqua et Novi

Here are some of the X accounts I follow because, frankly, they do cool stuff, and my immediate emotional response is all that matters, right?

(@)Macbaconai

(@)piotrbinkowski

(@)miboso__

(@)AllaAisling

(@)astronomerozge1

(@)BreezeChai

(@)pastpassengers