Beyond the Brush: Embracing Creativity with AI Art
It seems that everywhere you turn nowadays, someone is hating on AI; comparing chatbots to “deranged toddlers” and railing about how misguided it is to use the word “intelligence” to refer to it. For the most part, I skim past these, and go ahead using AI to do a lot of things that I was unable to do beforehand. I sometimes repeat the now-conventional wisdom, “You won’t lose your job to an AI. You’ll lose your job to a person who knows how to use AI,” and I am making sure that I’m a person who knows how to use AI, in lots of ways. I’ve documented some of them in this blog, and plan to document a lot more. I use AI just about every day. Arguments against AI don’t get me uptight, I just see a lot of people who don’t share my vision. That has happened a lot in my career, and will continue to do so.
But there is one argument about AI that does get me uptight. And I find it a bit surprising; why should I care what a bunch of people think about some particular aspect of AI? And which aspect is that? It’s AI art.
The objections to AI art are pretty well known. It is a “plagiarism engine”, that takes “mountains of stolen, unsourced art, fed like a grinder into a sophisticated algorithm that is, fundamentally, uncreative.” The courts have taken up the plagiarism case, challenging the plaintiffs to actually come up with someone who has standing to press charges; none of the AI images bear much resemblance to any particular work of art. This opinion might change, of course; that’s why we have courts of law, to systematically pursue questions like this.
I admit I have become addicted to making AI images — I won’t call them “art” since I really don’t care if they are art. I have commissioned a number of pieces of original art from artists over the years, and purchased a number of others. None of these are investment art; I don’t think they have any resale value at all, not to mention, any increase in value. I bought/commissioned them because I like how they look, and I have them on display in my home.
Are the works I have purchased “creative”? Are they even “art”? I commissioned one artist to take a photo I had taken, and transcribe it into a pen-and-ink drawing in his particular style. For all I know, he just projected the image onto a canvas and started tracing. Or maybe he did something a lot more involved. I don’t really know nor care; I am very happy with the final image, and feel it was worth the price. From the point of view of a consumer and patron of the artist, the method by which he created it doesn’t have any impact on my appreciation of the art.
So why am I uptight about the people who rail against AI image generation? It’s because of how empowering I find this technology. I am a mathematician by training; there are some very technical skills that I have that are pretty rare. And like many people with rare skills, they bewilder me; I really can’t fathom how someone can find math difficult; if you want to understand something that you find difficult, just concentrate some more. Study some more. I just can’t imagine how someone can say that it is just too difficult, and they just can’t do it, no matter how hard they try. But in fact, there are a lot of people like that.
But I can’t do that with drawing. I can imagine an image I want, but I just can’t for the life of me figure out how to put pen or pencil or brush to paper or canvas to make it happen. And I really don’t think that studying and drilling will help; I just don’t have that talent. I think of this when I try to understand how someone can just ‘not get’ math.
I don’t think my experience with the decorative art around my home is particularly unusual, nor my experience with commissioned art works. I give the artist a vision, and they execute it, bringing something of their own to it. There is an opportunity for some give-and-take, but for the most part, the artwork is finished, there’s no second shot. The work is a collaboration between patron and artist, but there’s just no getting around it, the artist gets the final say.
But with AI image generation, there is a lot that the operator can say. I can make literally hundreds of drafts of any image, and make a selection of one that I think is worth printing and hanging in my home. But here’s the most important thing: more than any commissioned art work, more than any image I find on the web, more than any stock photography, I feel that I have played a key role in the creation of this image. This is an image that nobody else on earth has ever seen, and it’s one that I participated in creating. With AI image generation, I get the last word.
Not all of the images go on my walls (I don’t have that many walls). Some of them go onto greeting cards. I have a friend who is very fond of cats and who likes to gamble; so I wanted a birthday card of kittens playing at a roulette wheel. There are a lot of images out there, but I haven’t found one that is suitable for a birthday card. And even if I did, that certainly would be stealing someone’s art work (or maybe, just using someone’s daily snapshot in a very odd way). But with AI image generation, I can control the aspect ratio, the composition, and come up with a personalized birthday card.
I realized why the AI image haters bother me. When I create images like these, I feel like an artist. I take ideas I have in my head, and I coax some recalcitrant software into creating an image that satisfies my imagination. I realize that when someone attacks AI image generation (and I’m careful not to call it “art”; just like most people don’t think of themselves as mathematicians, I don’t think of myself as an artist), they are saying that I don’t have a right to feel this way. Surely this isn’t the intention of the critiques, but it seems to me that they are saying that I don’t have a right to to express my own creativity. After all, I’m a mathematician; I have no imagination, no vision, no creativity. And while we’re at it, the woman who runs the till at my grocery store, and the barista who makes your coffee, and Hazel in Accounting. None of these people have imagination or creativity. Since we either don’t have the talent, or maybe the time and freedom to develop the talent, we don’t have imagination.
Here’s another example, which I don’t have to write up, since one of the friends on my Christmas Card list has already done so. I sent a card to Arnold Zwicky, who told the whole story on his blog. I really encourage you to stop reading here, and go read Arnold’s blog, he’s a delightful and entertaining writer. Now that you’re back, you can see how he reacted to this image. Is it “art?” — I really don’t care, if I can get that kind of reaction from an old friend. I also encourage you to search Arnold’s blog for his posts about some other cards I sent him; search for “Christmas”. I will warn you, as will Arnold; some of his blog entries have adult themes, as did the cards I sent him.
I hope these stories have given you some idea of how personal the experience of using AI to create images can be. Have I ‘stolen’ from some artist? As far as I can tell, nobody has published any images of Christmas Unicorns playing vinyl records; I didn’t take their image, or even any recognizable part of it. Arnold identified a source image that probably informed the AI that built my image. It is a classic work of art from the 15th century; we don’t normally consider long-deceased classic artists from the middle ages as still having legal rights to their images. There are certainly more modern images that led to the training of the AI that generated these images. How much of the $0.00 that I got from making this card and giving it to Arnold should I share with them? Or maybe I should just credit them (if I could figure out who they are). I’m sure that no artist will have any issues having their name associated with the sort of images I sent Arnold after Christmas (you looked, didn’t you?). Am I taking a job away from another artist? For the after-Christmas cards, if not for AI, I wouldn’t have done this project at all; commissioning such a project is too time consuming and costly for a one-off action. How about the Christmas card? If I hadn’t used the AI, I would have just gone to Hallmark or some other shop and bought mass-produced Christmas cards. I’m not sure how much of the cost of that box of cards Hallmark would pass on to the artist, but I bet it’s not much. And I certainly wouldn’t have been written up in Arnold’s blog if I had. And in any case, the whole thing wouldn’t have been as personal. You can say that I’m not being creative all you like, but it is undeniable that the cards I create have a personal touch that is impossible from mass-produced commercial art.
I have noticed a difference in how I view art, since I have started spending a lot of time creating AI images. When I go to a local art fair, and I see the works on display there, I can’t help but find myself thinking, “I wonder what model they used for that” or “I wonder how they did the prompt for that”. Then I realize that I am looking at an artwork done using some other medium than AI; then I ask myself, could I have done this with AI? Sometimes the answer is “yes”, and other times the answer is “no”, and in the “no” case, I’m even starting to get a feeling for the “…because”. I have developed a more critical eye to art because of my experience with AI image generation.
I have become particularly interested in multi-media art; the sort of thing that remains very difficult for AI to do, since it involves manipulating something other than an image. And while AI offers me a personal connection to a piece of art, the same can be said for more traditional art, in a different way. Take the case of A J Heckman, a local artist here in Columbus, whose work is abstract and colorful. Exactly the sort of work that is often subject to the criticism that “That’s not art, I could do that!” Obviously, I disagree; I have two of AJ’s originals hanging in my home. They have a story; there’s a bicycle race every year here in Central Ohio called Pelotonia; it is a benefit event for the cancer center. AJ set up a station along the route, with some paint and some old bicycle wheels. He invited participants during their rest break to dip a wheel in a color and spin it over the canvas. I suppose AJ’s detractors are right, anyone could have done that; but AJ is the one who actually did it. I saw his work on display when I took my mother to see a show at Shadowbox Live; the theme of the show was cancer. AJ’s Pelotonia work was on display. I purchased both canvases, which now hang in my home. My mother passed away from complications due to cancer about a year later. These works have a special connection that goes beyond just the images themselves. My experience with AI image generation has made me more aware of this aspect of art; how it can connect beyond the work itself. Simultaneously, it has made me more critical of art that fails to have such a connection. I have become a more informed art consumer.
Back in the 19th century, the technology of photography was met with similar skepticism by artists and art critics of the day. But when the smoke cleared, what we realized was that visual art wasn’t all about realism. In fact, it might not be about realism at all. This awareness brought about movements in art like Impressionism and everything that followed; some of the greatest works of visual art came from this revolution. Eventually, photography was recognized as an art form in its own right, as it became clear that there was a lot more to artistic photography than just “pointing a box at something and pushing a button”; just as there’s a lot more to AI image generation than just “writing out a few words and letting the machine do the rest.”
AI is a tool that lets humans of all walks of life feel like artists; to experience, maybe only a little bit, of the charge of creating something they (and their friends) feel is beautiful.
