AI Art is Not Art
Art is Human
I recently read an article from Aesthetic for Birds which makes the interesting claim that AI art is art. Although the article certainly raises an interesting defence of AI art, I find it to be unsatisfying in final review. I would like to raise the argument that AI art is not art at all.
What is Art?
So how do we define art? Art can be defined in a myriad of ways, but I propose there are three key elements. First, art is defined by the sensory experience of the person viewing it. Second, art is expressed through the creativity of the artist. Third, art serves as an interpretation of the world — it is trying to make a point.
(1) Art as Sensory: There needs to be something there to experience (or lack of expected experience) for there to be an art piece. Whether we see, hear, smell, taste, or feel it, for something to be art it must be experienced by the senses.
(2) Art as Creativity: Art serves as an expression of creativity by the one creating it. For something to be art, it needs to be created, serving as an expression of the creative function of the person bringing the piece form.
(3) Art as Interpretation: More controversially, art requires a function as the interpreter of the world. If we did not define a function of art, anything anyone does every could be defined as “art” and the term loses all meaning. Rather, art should reveal something about the world. If this sounds broad, that is because it is.
Using these three elements, I argue that we can define art as something that was intentionally created to provide a sensory experience for the sake of interpreting the world.
A critical perspective is Roland Barthes’ claim that the birth of the reader means the death of the author — or that the act of interpretation does not (and in fact denies) need authorial intent. I would agree with Barthes that the subjective interpretation of art does not require the author, however I would add the caveat that the author is a uniquely positioned viewer with respect to the piece. Therefore, I believe there are two levels to applying the definition of art I propose: the level of author(s) and the level of viewer(s). The author can view a piece as art, as they know their intentionality and interpretation, but a viewer can deny that they see any meaning in the same piece. This means that something can be seen as art and not-art simultaneously.
In this conflict between the levels of interpretation, can we place a priority between the author and viewer? Following Barthes critique, we see that the author clearly cannot be prioritized over the viewer. And who is to say that any particular viewer is superior to the other? Hence, I argue that no reference point is superior, and we must retain the contention of art-not-art that exists between author and viewer.
Another interesting conclusion to the simultaneity of art-not-art is that our ability to derive meaning from a piece changes dynamically with time. There was once a time in my life where I did not see any meaning in art by people like Joseph Havel (see the piece above), but now derive much from it. This means to the viewer, not-art can become art (or even more interestingly, art can become not-art as it loses meaning — like Kurt Cobain and Smells Like Teen Spirit). Even on the level of author, whether something is art can change dynamically. Additionally, these levels are not as discrete as it would seem at first. The viewer’s opinion can change dynamically based on knowledge of the level of the author (and vice-verse for the author). We see then that not only does defining something as art change based on level in a static sense, but that even on a particular level, the definition changes over time. All pieces then sit in the limbo of art-not-art.
What do we say is art, then? Nothing? Everything? In fact, what art is is what you think art is.
So What is Art?
Let’s then apply this definition of art as interpretative intentional sensory experience. Most people would say that a piece by Michaelango, or van Gogh, or Caravaggio are clearly art. What about postmodern art? Many people do not find any meaning in them, and interestingly, some authors of these pieces claim there is no meaning in their own pieces. I would say, on those levels, then there is no art to be found. However, I personally would argue that in my perspective that postmodern art is in fact art as I derive meaning from it.
Can art be a raw display of talent or beauty in of itself? I would argue that no, it is not necessarily art. If the viewer is moved by the beauty they see in a piece, then on their level it is clearly. But if the painting they are moved by is a photorealistic painting of a bowl of fruit that simply serves as technically admirable, I would say there is no art on the level of the author. The author has not imbued any meaning into the piece, and hence it does not serve as art, regardless of its precision. This also helps us discern the difference between someone taking a picture vs photographic art. The difference is not quality or effort, but rather the intentionality in revealing meaning through the medium. In this sense, we can say that a family photo album is art, whereas wildlife photography is not necessarily.
Let's consider the dynamic aspect of viewing a piece as art. Let’s say that a viewer is looking at a Dalí, and then comes to the knowledge that it is forged. They will likely no longer view the piece as art, but counterfeit. On the level of author, of course, the piece is not art because it is not an intentional act of creativity, they merely copied it. The piece starts as not-art from the author, is temporarily art for the viewer, but then eventually returns to being not-art due to the instability caused by the level of author.
Based on this reasoning, I would argue that AI art is not art on the level of the author because there is no intentionality or creativity behind the design. Algorithms are unable to make decisions as they do not have minds, and hence it violates the definition of art. The viewer, of course, can consider the piece to be art themselves if they find meaning in them. However, there is a deeper issue with AI art not being art on the authorial level — it destabilizes the meaning of the piece altogether and transgresses the boundary between the levels. Not only is there is no creativity, there cannot be any interpretative meaning within the piece. Dynamically, for the viewer to gain the knowledge that the piece was AI generated actively reduces the ability to see the piece as art for many viewers. In a strange inversion of Barthes, the author being dead ironically is the death of the viewer.
The lack of any sense of the level of the author seems to destabilize AI art as art altogether for the viewer. If one were to derive meaning beyond this point, it would be an absurdist act against the void found. AI art only exists in the hyperreal — there is no wizard when we pull back the curtain.
Art as Human
There is something very human about art. We express our creativity with one another, we experience it as fleshy embodied people, it causes us to think about the world in ways we never considered. We can be taken aback by its beauty, we can be moved to tears by its pain. To move art away from the embodied human experience to a generic methodology of production that can be replicated by strings of code is to devalue a fundamental part of what it means to be human. Of course, we can find beauty in the pieces generated by these algorithms, but to move art completely away from the human person is to move away from art itself. Even if we move away from an anthropocentric view and talk of animals creating art, it is still an embodied experience of a living being. Art serves as the embodied integration of our minds and bodies, the interface between thinking, feeling, and sensing. There is no art for the brain floating in a vat. A computer cannot create art for another computer, it has no ability to think, to feel, or to sense. AI art cannot be art because it is empty.
