With the rapid development of “artificial intelligence” in recent years, we have come to see the emergence of so-called “AI artworks.” For instance, AI can now be used to generate images resembling a grand painting in the style of Rembrandt, or even a somber song like that of Miles Davis. In light of this technological development, one might come to question, or even outright dismiss, the long debated ideas of aesthetic philosophy that were once concretely a part of its dialectic, such as whether art is the work of a human, or, further, if beautiful art requires something like a savant or “genius” (as is Kant’s view) for its creation. If AI gets to the point where it can generate something indistinguishable from a work of Picasso or O’Keeffe, what even is the point of artists anymore? This essay will utilize Kant to help reconcile the status of “AI art,” and seek to demonstrate how, at best, AI alone lacks the “spirit,” that is, for Kant, the “attunement” to the ultimate taste of nature, which is necessary in creating fine art. Further, we will look at Dorothea Winter’s essay, “An Aesthetic-Philosophical Analysis of Whether AI Can Create Art,” which argues, from a Kantian perspective, that the primary distinction between art made by humans and by AI is the artist’s freedom. Finally, with Wassily Kandinsky’s work Concerning the Spiritual in Art, I will argue that art is only art if it has at least some spirit, responding to Winter’s essay, as artistic creation comes not merely from the free agent, but from the free agent with the capacity to express their aesthetic idea via spirit.
In Kant’s Critique of Judgement, he describes “art in general” as “production through freedom.” Art is not merely the “effect” of various causes (as if the brush strokes are “effects” of a painter), rather the artwork is the product of an artist’s freedom. The freedom of art in general is what distinguishes it from science and handicraft, as science concerns only the “sufficient knowledge of the desired result;” for instance, a chemist mixing vials is testing, and then theorizing about, the effects of their mixture, and not creating a product “in itself” for some deeper purpose other than to explain why the effect occurred. With handicraft, such as a shoemaker, for instance, the craft is conditioned by the needs of the business (to make a profit and to minimize expenses), and the shoes themselves are conditioned by their concept (that is, the worker must make a “good shoe,” and conform to the concept of “shoe” to sell them). Art then is distinct by its “freedom,” as a work of art is not a means to communicate a rational idea (like with handicraft and science), but is the end goal itself for the artist. This distances art further from political cartoons or advertisements, which may utilize the techniques of art (i.e., in its design, symmetry, or color theory), but always for the fulfilment of some external end (i.e., to convince the audience to give them money or vote).
One may doubt Kant’s claim here, as plenty of genuine artworks depict various vast ideas, be it God, good overcoming evil, heaven and hell, or some other “ethereal” concept, like with Michaelangelo’s Creation of Adam; in these cases, however, what we see is not merely the intellectual representation of rational ideas (as if Michaelangelo, looking to paint God, had read all the works of Thomas Aquinas so to accurately depict the concept of God), but rather the representation of aesthetic ideas. Like rational ideas, aesthetic ideas “strain after something lying beyond the confines of experience,” however, where rational ideas rely on logical inferences of language (like with syllogisms), “no concept can be adequate to [aesthetic ideas] as internal intuitions.” Kant describes aesthetic ideas as “the counterpart (pendant)” to rational ones; “pendant” implies a complementary relationship between the two, and clarifies the aforementioned doubt:
[The artist gives] sensible form to the rational ideas of invisible beings… transgressing the limits of experience he attempts with the aid of an imagination, which in reaching for a maximum emulates the precedent of reason, to present them for the senses with a completeness of which nature affords no parallel.
Kant understands art as the representation of aesthetic ideas (internal intuitions), which are, by their essence, expressible not by concepts, but only through a work of art; the artist can utilize various concepts, and, via imagination and understanding (the artistic “emulation” of rational concepts), make sensible what is otherwise totally beyond experience.
Though it is now clear that the artist employs aesthetic ideas (and not rational ones), both the origin of these ideas from within the artist, and the faculty which makes their presentation possible, have yet to be clarified. Before doing so, we will take a moment to look at Winter’s essay, as her solid explanation of the freedom necessary for art will open the possibility for us to introduce “spirit” into the equation. In the third section of her essay, titled “According to the [Kantian Approach to Art], There Is No Art Without (Artistic) Freedom,” she explains how Kant “demands the use of reason” of the artist, which is to say that the artist must employ both téchne and episteme alongside their free expression of their aesthetic ideas. Winter understands freedom to be predicated by the artist’s “[liberation] from the conditions of the moment,” which allows artists to “extend their creative ambitions throughout time,” so as to give as much time as necessary to the artwork; only once the artist is free in this way is the creatively free expression of aesthetic ideas possible. Freedom further entails an approach to making art that is essentially divorced both from its public perception (that is, the artwork is not made with the purpose of pleasing everyone) and whatever monetary value it may hold.
While it is clear that art, for Kant, is contingent on its creator’s freedom, it is unclear how the artwork manifests freely from the artist: sure, the free (from temporal or monetary incentives) expression of aesthetic ideas is that manifestation, but the origin of these ideas thus far remains a mystery. Is the artist merely picturing some “artwork-yet-to-be” that no one else has thought before, and then creating that new thing? The genesis of the aesthetic idea is perhaps the most important element of Kant’s philosophy of art, and for the most part, this is left out of Winter’s analysis. She does note, however, the importance of the subject’s spontaneity in the art creation, writing that “Kant argues that human beings are capable of creatio ex nihilo (i.e., creation out of nothing) because human beings are able to act autonomously.” She argues that this is the main difference between Human and AI generated art, as the images an AI generates are “ultimately a form of permutation,” regardless of the quality of the generation. I do not wish to say that Winter’s analysis is wrong in asserting human freedom as a definitive element for the creation of art, but this misses the complexity of Kant’s thinking about art generally and the artist. Total Freedom, more specifically, is necessary for fine art; while all art that humans make is a product of freedom, there is still “art” that humans make that is not totally free.
This is evident Kant’s analysis of “spirit:” firstly, Kant frequently critiques art as “aping” throughout his analysis—“aping” is the mimicry of other artworks through the copying of their “method,” such as someone, in an effort to make a great work of art, methodologically copying Van Gogh’s style of brush strokes or color. In an opposite sense, “mannerism” is the intentional “distancing from imitators,” where someone makes something with the desire to be wholly original. Kant notes that mannerism is still necessary for the creation of fine art, as manner “possesses no standard other than the feeling of unity” in the artwork. Where pure mannerism fails to be fine art is where it lacks spirit: Kant explains that “[the styles] intended to mark one out from the crowd (though spirit is wanting), resemble the behaviour of a man who… were on a stage to be gaped at.” The difference between a derivative work of art and an authentic expression of aesthetic ideas is not that the latter is free and the former is not (as both works are doubtlessly created by a free agent in either case, though the level of the employed freedom can be said to be different, as the aping work is derivative), but that the authentic work is made with spirit. Spirit is necessary for the distinction of human and AI art because as long as human freedom lacks the component of spirit, the genesis of aesthetic ideas comes arbitrarily to the artist, and the origin of the work’s aesthetic idea cannot then be meaningfully distinguished from AI generated images.
Kant understands spirit as “the animating principle of the soul;” “a poem may be pretty and elegant, but is devoid of spirit.” Spirit is a type of “attunement” to nature which gives the “rules” to artists; Kant opts to prioritize the “genius” over the artist in this sense, as the genius is someone with a “talent” for creating art (where talent means a profound spiritual attunement). Spirit is the reason for art “schools” (baroque, neoclassical, cubism, etc.) and artistic lineages: Picasso, for instance, was “inspired” by Paul Cézanne, and he did not imitate Cézanne’s style, rather, discovered his own through experiencing Cézanne’s works. “Inspiration” shares the Latin root “spiritus,” which is where we get the word spirit; in this sense, we can think of spirit as the translation of the manner of expressing aesthetic ideas. I use “translation” here to evoke the way a linguistic translation simultaneously leaves out and adds various meanings to the original un-translated phrase. For example, we can translate both of the spanish words “conocer” and “saber” as “to know,” yet, in Spanish, both of these words have distinct meanings that are not captured by the English “to know,” as one could say “ya conocí Miguel” to mean “I already met Miguel.” Translation in regards to spirit can be thought of as the way Picasso “deviates from the common rule[s]” of Cézanne, and all prior artworks, in expressing his aesthetic idea, yet remaining faithful to the common beauty captured by Cézanne (which originates from the beauty of nature). This is why Kant writes that “the product of a genius… is an example, not to be imitated (for that would mean the loss of… the very spirit of the work), but to be followed by another genius.” The artist “following” prior works is the spiritual translation.
This idea gains potency when we contextualize it with Wassily Kandinsky’s concept of “spirituality.” Kandinsky was a painter well known for his abstract approach, and perhaps provides the most intimate perspective into the role of spirit in art creation. While there is no evidence that Kandinsky ever read Kant’s third critique, in his work Concerning the Spiritual in Art, we find many similarities between the two thinkers. For one, to return to the aforementioned idea of “levels” of freedom in art, Kandinsky provides us with the diagram of the “triangle” for thinking of the “levels of spirit” in art. He asks us to picture “a large acute-angled triangle divided horizontally” into three unequal segments; the lowest segment is the largest, then the middle, and the smallest third at the tip of the triangle:
The triangle is also moving, constantly, “almost invisibly forwards and upwards.” From a Kantian perspective, the genius alone would be at the apex of the triangle; Kandinsky gives the example of Beethoven and his cultural reception, who stood “solitary and insulted” throughout his life, and “abused as a charlatan or madman.” As time passed, and the triangle moved, Beethoven entered the lower segments, and the beauty of his work became recognized as such by the masses. Conversely, some artists may create a work in line with the middle section or the base of the triangle, and their work may be “fully understood by their fellows and acclaimed for their genius.” We can understand the bottom sections as composed of the “popular arts,” like the contemporary genre of “pop-music,” which is named so due to its popular accessibility. All the arts made anywhere in the triangle are products of freedom insofar as they are artworks, but we can also think of the triangle as depicting levels of conformity to a common rule, where the bottom has the widest range of appeal, and the apex with the least, though conforming to a higher taste beyond our culture’s understanding (like how Kant understands the genius to be given the rule by nature).
While Kant understands the genius to be the sole factor that “moves art forward,” as only through their faculty of genius can the artist come to deviate from the rules in a tasteful way, Kandinsky specifies that within each segment of the triangle are artists who move it forward. He explains that “each [of the artists] who can see beyond the limits of his segment is a prophet to those about him, and helps the advance of the obstinate whole.” Kandinsky also takes a similar approach to aping as Kant, as he writes that “efforts to revive the art-principles of the past will at best produce an art that is still-born.” Kandinsky understands spirit to be a type of “vitality” of expression, and that to emulate past works of art (like Greek statues) is to create an “unlived” work of art; the experience of the world that made the creation of Greek statues possible in the first place is long gone, which is why Kandinsky goes on to call this imitation “soulless,” despite any similarity it may have with the originals. To sum up, Kandinsky takes the triangle to represent the movement of spirituality in the world, manifesting itself through artistic expression, yet each layer contributes to this movement.
Before we go further into Kandinsky’s critique of the contemporary reduction of art into materialism, we will briefly lay out what exactly is meant by “AI,” as then we can consider not only the status of AI art (whether it is (or can be) or not), but also the role, as in, what influence it has over the triangle of spirit. Winter explains that “AI-generated art is an umbrella term that includes any form of art that cannot be generated without the use of programming.” While there are many advanced AI programs that can generate images (DALL-E; Sora; DeepDream), to constrict ourselves to one particular technology would be shortsighted, as within the last five years alone we’ve seen exponential development of what AI can achieve. That said, what unites them all is that a). They require programming to determine how to generate any image, and b). They require a set of images to reference in their generation. Winter goes on to explain that there are AI’s that do not need a prompt to generate images, unlike Chat GPT, and that they can instead input a vast collection of different works which the AI can then “learn” to generate certain aesthetic styles or art movements. As winter summarizes, “from a philosophic-logical standpoint, [AI generated images] are fundamentally if-then relationships,” as they must reference their database and put different elements together following its programming.
The fact that AI can generate images “unpromptedly” is not evidence for the fact that AI is creating something “ex nihilo,” as Winter stated earlier. She writes that “when DALL-E is reduced to its aesthetic-philosophical principles, the input-providing human remains the artist, regardless of the complexity of the algorithm or the excellence of the outcome.” All this is to say that while AI generated images could be “constituted” as art by its artist, AI is not the artist insofar as it is not free in its image generation. If AI art, however, is using a database of artworks to pull from in its generation, the most apparent difference between that and the work of a human is precisely in that the AI is not following the spirit of the artworks, as human artists do, but following its code. Winter explains “freedom must not be bound to calculability. It is the essence of freedom to remain free of computational necessities, as Kant points out in basing the act of creation on a rationally incomprehensible dialectic.” (The “rationally incomprehensible dialectic” being the fusion of imagination and understanding in expressing aesthetic ideas.)
One element of Kandinsky that we do not find so explicitly in Kant (outside of his critique of aping) is the reduction of art to materialism. Currently, the largest section of the triangle holds “materialism:” not limited to art, materialism “has turned the life of the universe into an evil, useless game.” With the 20th century, “scientism” has grown to become the dominant mode of thinking, where what “matters” to people is merely the material objects. Kandinsky also critiques artistic “realism” in this way, as its “aim was to reproduce anything ‘as it really is’ and without fantastic imagination.” Is this not precisely what Kant would see as “soulless art,” as a realist approach would work to essentially distance its creation from the artist's imagination? This also suggests that AI art, constituted so by an artist, would have to be drastically different from the works which it pulls from to contain any spirit at all; with the 20th century “dada” movement and “ready-made art,” artists took ordinary objects and asked the audience to view them as the work of art itself. For example, Marcel Duchamp’s (in)famous Fountain is merely a urinal laid down (with “R. Mutt” written on the side). Whether ready-made art is or is not art from the Kantian perspective is not the discussion here, rather, if, at the very least, we take Fountain to be Duchamp’s expression of his aesthetic idea, we can see that even here is the “following of spirit,” as ready-made art is a particular response to the history of art, and not an arbitrary work. An AI image, as it is replicating the works in its database, is a type of “realism,” if not a simulacra, of art generally: for an AI generation to be considered art in an authentically spiritual way (and given that it is only constituted as art by its artist), it would have to generate something without any database, or maybe without a database of artworks, as the generation would be working to replicate that data as accurately as possible (which would give us nothing more than an image that merely resembles a “real” work of art).
Winter writes that “as an artistic medium, AI opens new and expanded possibilities for human artists and art as a whole… AI will be increasingly relevant to the field of art in the future.” This is true only if the AI art has spirit: following Kandinsky’s critique of materialism, utilizing AI in creating art holds the risk of reducing art to realism. When art becomes something that must only resemble art, we lose touch with the spirit that drives artists to create in the first place. Kandinsky anticipated this reduction and hence proposed a “spiritual revolution”—if we want to salvage the possibility of creating art via AI in a meaningful way, it can only be through something akin to an abstract approach. As Michel Henry explains in his book on Kandinsky, Seeing the Invisible, “abstract art offers the new appearance of forms attached to nothing… It is painting without obligation or sanction.” Abstract art is perhaps the highest form of fine art in the Kantian sense, as it is the most free expression of aesthetic ideas, with an essential distance from realism (as no concept is particularly depicted, rather the “feeling”).
So, if AI can make something abstract, then what is even the point of painters anymore? Are painters not then taking the least efficient road to their expression if a work can just as well be produced with an AI in only a few minutes? For Kandinsky, AI would never be able to match the “inner” element of expression found in a human made painting, primarily due to his idea of “composition” and “harmony.” Composition does not mean solely the arrangement of the work in a technical sense, but the creation itself (which is, to use Kant’s language, the technique alongside the aesthetic idea). We can think of “harmony” similar to Kant’s notion of the “rationally incomprehensible dialectic” in creating art, as he explains:
The composition arising from this harmony is a mingling of colour and form each with its separate existence, but each blended into a common life which is called a picture by the force of inner need.
For Kandinsky, both the expression and the reception of the work of art is an intrinsically subjective relationship between the subject and the work: this is to say that art necessitates, from the viewer and its creator, a lived experience. The artist, in expressing their aesthetic ideas, is pulling from their own spiritual engagement of the world; spirit here means the emotion that is inseparable from the work, and can only come from having experienced that emotion—to make art is a transcendental process of sorts. AI has no lived experience, and, as Winter described it as an “if-then” process of creation (this also applies if AI were to “judge” a painting), AI has nothing to express through the technical elements of making art, unlike a human artist with a receptivity to the world. AI then has only the “subsidiary” elements to its generation, but no inner substance (as in, no meaningful arrangement of color or form, besides its closeness to the works of its database). This is why the title of this essay is “the critical agents,” as the viewer and the artist have a necessarily non-reducible attunement to spirit that ties Kant’s whole aesthetic philosophy together, that is, the subject’s capacity to recognize beauty and taste in art, and also the faculty of genius in imagining aesthetic ideas that move art forward. Humans, free and spiritual, are critical to the creation and existence of art. Insofar as the beauty of art is a representation of the beauty of nature, for Kant, AI art that mimics (“apes”) human art is a type of “trick” that we play when presenting it alongside human works of art, as he explains that:
Were we to play a trick on our lover of the beautiful, and plant in the ground artificial flowers… and were he to find out how he had been deceived, the immediate interest which these things previously had would at once vanish.
While beauty is a “disinterested judgement” for Kant, our capacity for taste creates an intellectual interest in the beautiful, and to disguise AI generated art (that is not abstract, and merely mimics art) suppresses the spiritual importance of art, which would lead us to something like a Kandinskian materialist nihilism.
While AI technology is indubitably impressive, our fascination runs the risk of reducing art into something that merely looks like art, and lacks the essential parts that make art meaningful. Of course, there is the possibility that AI can be used in artistic ways, but it would require a complete restructuring of our approach to art in general so as not to reduce it to its material; the reduction of art is not limited to AI, which is how Kandinsky was able to write about the dangers of prioritizing realism. While Kant never anticipated AI itself, nor the impact on art that AI would have, we nonetheless find “spirit” as the most powerful element of his aesthetic philosophy when it comes to what makes art meaningful, and simultaneously differentiates soulless AI art from human art, while also allowing the possibility for the triangle to move forward alongside the future of AI.
Citations
Kant, Immanuel. 2008. Critique of Judgement. Edited by Nicholas Walker. Translated by James C. Meredith. Oxford University Press.
Winter, D. (2024). Aesthetic Aspects of Digital Humanism: An Aesthetic-Philosophical Analysis of Whether AI Can Create Art. In: Werthner, H., et al. Introduction to Digital Humanism. Springer, Cham. https://doi.org/10.1007/978-3-031-45304-5_14
Kandinsky, Wassily. 1977. Concerning the Spiritual in Art. Translated by M. T. H. Sadler. Dover Publications.
Henry, Michel. 2009. Seeing the Invisible, On Kandinsky. Translated by Scott Davidson. Continuum International Publishing Group.
No comments:
Post a Comment