Artificial Intelligence (AI) is all the rage these days — that and aliens. While I admit the topic is not of great interest to me, all the chatter prompted me to explore this new technology — dialoguing with ChatGPT to test its capabilities, talking with programmers, and so on. Having resurfaced from this rabbit hole, I thought I would offer some thoughts on AI that draw on my areas of expertise. I expect this will be the first of several posts on the subject.
To all my subscribers, thank you for subscribing. To my paid subscribers, thank you for your support. And to any visitors, please consider subscribing and supporting my work. Enjoy!
I begin with this. Can AI create art?
Readers who know me solely as a scholar of philosophy and religion may wonder what qualifies me to speak on art. Suffice it to say my study of philosophy and religion was preceded by formal training in the fine arts, training I used in a professional capacity during my graduate studies and still use today. To avoid a lengthy aside, I will relegate the details to a note.1
In my exploration of AI, I have come across a good deal of AI “art.” To name just a few examples, someone shared with me AI renderings of the book of Revelation, another showed me AI concept art for a television show, and when searching for an H. R. Giger piece, I stumbled upon a series of AI mimics. I do not deny that many of the images are impressive despite obvious defects. — Why does that horse have two heads and six legs? — But I have no doubt these wrinkles will soon be ironed out, and I also have no doubt that such images signal the impending death of a great many jobs. Why pay a concept artist when AI can do the job quicker and cheaper? (I will resist offering thoughts on this particular question until a later post.)
The question at hand is not whether AI can generate images that look like the work of an artist. It certainly can. Nor is it about the comparative quality of those images. Fun though it is to joke at the ways AI struggles with things like hands, the fact remains that it is capable of effectively mimicking various artistic styles. In fact, I admit that the AI mimics of H. R. Giger are technically superior to Giger’s own work. Giger excels at producing nightmarish images, but his work betrays technical deficiencies in his ability to render the human — and thus humanoid — form. The AI mimics capture his nightmarish touch without these technical detractors.
The question at hand is whether the images AI generates are, in fact, art. I submit the answer is No. AI does not and cannot generate art.
To see why, we must define the term “art,” along with related terms, such as “artist” and “artwork.” So, let’s begin there. Art. What is it?
Many today presume art is indefinable. Modern “Art” has so deconstructed the word that it leaves the average person at a loss. Apparently, ordinary objects can be declared art, and there is no good or bad about it — Marcel Duchamp’s “Fountain” (or signed urinal) being the iconic example.
So, is this true? Does the term “art” signal a purely subjective social construct — a mere label that one can slap on whatever he pleases? Or does the term refer to something real, an objective feature of reality?
The question is pivotal. For if the word “art” is an empty noise that one can apply to anything willy-nilly, then the question of whether AI can create art has no objective answer, only subjective ones. Some may label AI images “art” while others refuse the label, but neither choice is more true than the other, for there is no reality behind the label itself. If, however, “art” refers to something real, if the label identifies a substantive feature of reality and, thus, can be correctly or incorrectly applied, then we have grounds for determining whether or not AI is producing art.
Some readers may recognize this fork in the road. The fork about art — whether it is an empty name or something real — is the same fork in metaphysics that I have discussed ad nauseam, namely, the fork between realism and nominalism. I will not belabor the distinction, but the present topic merits a brief refresher.
The realism-nominalism divide recognizes that the mind naturally looks out into the world and sees organizational structures of genus, species, and common properties — for example, animal, human, bipedal. The question of realism is whether the mind identifies structures that are part of the world and abstracts these structures from the world itself or whether the mind invents organizational structures and imposes them on a world void of such structures. The former position is called “realism” because it holds that these structures are real. The latter is called “nominalism” — from the Latin nomen, for “name” — because it holds that these structures are mental fictions, mere names imposed on the world.
Any who know my work knows that I am a realist. And this extends to my view of art. I hold that “art” signifies something real, and the label can, therefore, be applied correctly or incorrectly. Art can be good or bad in quality. Various art forms have various teloi — ends or purposes — that elevate some art forms over others. And the list of objective truths about art goes on. Of course, not every person judges rightly whether something is or is not art, or whether the art is good or bad, beautiful or ugly. Such is the distinction between epistemology (knowing) and ontology (being). But that there is such a thing as art which bears such properties is true. And for this reason, we can meaningfully assess whether or not what AI produces is art.
I will begin by laying bare my philosophy of art.2 So, again, what is art?
The etymology of the term draws us to the Latin ars, which refers to a cultivated skill. Let’s take as our example the skill, or art, of drawing. As a rational animal, the human person comes into being with the potential to cultivate this skill. Like all creatures, the human being undergoes a process of generation, moving from potentially this or that into concretely this or that. In the womb, the material potential for heart, brain, legs, and arms moves into concrete reality as matter manifests the human nature embedded within it. A great many of these potentialities naturally grow to fullness without any conscious effort on our part. But not all potentialities are this way. The potential to learn to speak or read or play the piano require volitional acts of cultivation. And so it is with a skill such as drawing. As a rational animal, the potential to draw is within me and within you, underwritten by the pairing of reason with a body capable of such a craft. The skill must be cultivated, however, if it is ever to advance beyond childish doodles.
The ramifications are straightforward. “Art” refers simply to the specified skill. In the case of drawing, this skill is the “art.” The meaning of “artist” naturally follows — one who has cultivated this skill. And when the artist articulates this skill in the production of a particular drawing, the artist produces an “artwork” or “artifact.”
This basic scheme helps dispel the fog of Modern “Art.” Despite rumors to the contrary, not everyone is an artist, and one does not become an artist by fiat. Only one who has cultivated an art is, in reality, an artist. The one who has not cultivated an art but declares himself an artist speaks falsely — no matter how flamboyant he may be in his declaration. Likewise, not every object that one declares art is, in fact, a work of art. To constitute a work of art, the object must proceed from an artist as an articulation of the art he has cultivated.3 To declare something art that falls short of this standard is to speak falsely.
I will forego discussing how this philosophy of art unfolds into an assessment of its quality, though I trust several ramifications are obvious. The greater the skill articulated, the greater its artifact. Art, as something real, has teleology, and some ends are higher than others, naturally elevating some arts above others. Many arts are susceptible to moral properties, which can further elevate some art by its moral goods or degrade art by its immorality. These and other facts of reality are all relevant to the assessment of the quality of artwork.
However, the question we are exploring here is not the quality of AI “artwork” but whether what AI produces is, in fact, art.
I trust the answer is obvious in the light of the foregoing. AI does not bear the potential to cultivate a skill. Hence, its capacity to produce an image is not art, nor is the machine itself an artist, nor are its images artwork. AI images are images, not art.
The first self-playing piano was invented circa 1895 by Edwin Votey. The device can play a piece of music flawlessly — that is, in perfect tempo without missing a note — but we immediately recognize that this precision is not the articulation of the machine’s skill. Yet, when a human being plays the same piece with the same level of precision, this act is an expression of skill, of a cultivated art. Hence, though the two songs may offer no auditory difference, the underlying ontology is different. One is art while the other is not. And the same is true for AI images.
I noted above that AI Giger mimics offer very similar nightmarish images to those produced by Giger himself but with none of the flaws in Giger’s skill set. So, is the AI image a superior work of art? The answer is No. For the AI image is not a work of art at all. Giger’s images are the concrete articulation of an art cultivated by the artist, Giger. The AI images are not. It is a manufactured image.
Ontology matters. Were there a machine — and there may well be one — that could scan Michelangelo’s Pieta and then chisel an exact replica out of marble, the two pieces would be identical — the machine-chiseled replica sharing every structural and material property with Michelangelo’s masterpiece. But the pieces would be substantially different because of a very real and essential relational property that belongs to only one of these sculptures. To wit, only one would be the handiwork of Michelangelo, and as such, only one would be the articulation of his unimpeachable skill set, of the art cultivated by one of the greatest artists to grace the face of the earth.
In short, though an observer would have trouble telling the sculptures apart, the fact would remain: One is art and the other is not.
I suspect that most of my readers know me either as a scholar of philosophy and religion (which I am) or as a film maker (which I am). Less common, I suspect, is the knowledge that I am a trained artist. I studied fine art painting and drawing at the Maryland Institute College of Art, which was, at the time, one of the top three art schools in the country. You can view samples of my work here. As a fine art artist, I’ve sold works to private collectors. In the commercial arts, I’ve worked as a concept artist, character designer, and pre-visualization artist in both the video game industry and then the film industry. As a professor, I’ve taught classes on aesthetics and art history. I’ve published on my philosophy of art and aesthetics generally and on the views of Sir Joshua Reynolds in specific. I participated in the framing of a grant program for the Templeton Religion Trust on “aesthetic cognitivism,” the theory that art and aesthetics may be capable of offering insights into truths that would otherwise be inaccessible. (You can find Templeton’s interview with me on the topic here.) I’ve appeared on the PBS series, Closer to Truth, discussing my philosophy of art and aesthetics, amongst other things. (You can find those interviews here.) All of this is to say that my thoughts on the philosophy of art are the thoughts of an artist who has applied his knowledge of philosophy and religion to his practices, reverse engineering a philosophy of art from the point of view of one who creates art.
Although a bit dated, I lay bare several features of my philosophy of art in a pair of articles in the Journal of Markets and Morality (volume 12, number 2, 2009), namely, “How Should Christians Be Stewards of Art?” and “How Should Christians Be Stewards of Art? A Response to Calvin Seerveld.”
Each component of this statement matters — the object must proceed from an artist as a concrete articulation of the art he has cultivated. If one cultivates an art and is therefore an artist, but he, like Duchamp, declares a disposed piece of trash to be art, he speaks falsely. His status as artist does not give him a magical power to elevate something that is not art t the status of art. Moreover, not every object that proceeds from him is in fact art; the object must proceed from him as a concrete articulation of his cultivated art. Anything less is not artwork.