On AI-generated art

 With the rapid recent improvement in the capabilities of art-generating AI algorithms, there has been a natural and predictable backlash against it from an outspoken majority of human artists.  This backlash has been largely financially motivated, and seemingly, from my perspective, driven in large part by professional freelance artists, particularly those who specialize in digital art, it seems to me, which is perhaps natural as the AI algorithms are generating digital art themselves, being of course unable to create physical paintings on canvas.

This reaction tends toward two arguments that I will argue are contradictory, overall making it difficult for me to find them terribly convincing in combination. The implied conclusion from these arguments is that AI generated art is bad and that we should stop using and publicizing these algorithms.  They are as follows:

1. AI generated art is inferior to art created by human artists
2. AI generated art is stealing from human artists and is immoral

I have seen Argument 2 more frequently, but Argument 1 is common enough to severely undermine Argument 2, serving as a persistent reminder of why I find myself if not un-sympathetic, then unconvinced, which I have found to be somewhat surprising given my appreciation for art and my natural inclination to support artists.  Two of my grandparents were artists, after all, and a distant relative of mine was Winslow Homer, who was a very influential figure in the history of American art.  Yet, from my perspective, none of these people would have been impacted at all by the development of AI-generated art if it had occurred in their lifetimes, for reasons I will argue.

Firstly, I will address the contradiction between the two arguments.  I actually agree with Argument 1, which makes Argument 2 much more difficult for me to agree with.  There seem to be many people who agree with both, which I find puzzling.  If AI generated art is inferior to human created art, then how does it threaten human artists?

The answer to this would seem to involve the following.  First off, Argument 2 must rely on idea that AI generated art, while to some extent inferior to human created art, is not inferior enough to make up for the ease with which someone can create AI-generated art.  In many ways, this is not untrue.  For someone like me, I would never be able to see a painting of, for example, Acadia National Park in the style of Tom Thomson, without spending lots of money to find an artist who could create such a painting, or myself becoming so proficient that I could do this myself.  The former would of course be much easier, but even that would be quite costly and time consuming.  What actually happened before AI-generated art was that this art (or image if the term art should not apply here) would never be created, and I would have to imagine what this would have looked like.  

This is the real issue with Argument 2.  The financial loss of opportunity created by an AI art algorithm is actually miniscule for the human artist in this scenario, which is the say, in the scenario in which quality is either important or unimportant.  Where quality is important, AI-generated art is insufficient to replace a talented human artist.  Where quality is unimportant, the scenario is generally one in which the consumer would not consider it worth paying for a human artist.  Where AI-generated art may deprive a human artist of work would seem to me to be scenarios where someone is willing to spend a small amount for something small and of moderate quality, comparable to a tattoo in size or complexity.  Such items may have enough sentimental value but be simple enough for consumer and artist to agree to a small fee, and perhaps, given that some of these algorithms can produce variations of images, these could be replaced with AI-generated art and deprive a human artist of work.

I suspect that most of the fear among artists generated by the proliferation of AI-generated images is largely speculative, and derives from a few sources.  Of course, prognostication of the impact of a development such as this is difficult to perform, and so it is not unnatural to attribute to AI art algorithms a level of quality that they do not possess, while simultaneously, out of defensiveness, deriding them as inferior to a human artist.  Such contradictory feelings are part of human nature, after all.  Many people in the art community have compared themselves (favorably, as I've seen) to Luddites, and it would not surprise me to learn that the Luddites believed both that textile mills would push them out of the market, and that they were inferior in quality.  They were not incorrect in this pair of beliefs, but the key difference in this case is that textiles are a necessary good, whereas art is not, and as a luxury good, has already been priced beyond the capacity of most people to afford it.  Luxury goods of this sort have continued to survive despite being undercut in price by cheaper competitors.  We can still buy fine lace and leather good, fine carpets, expensive watches and automobiles, and any other sort of luxury good despite there being cheaper competitors that are of acceptable quality.  The existence of a Honda Civic hasn't prevented Rolls Royce from continuing to exist.

I suspect there is also something of a conservative reaction, so to speak, to the development of AI-generated art.  Walter Benjamin, in his essay The Work of Art in the Age of Mechanical Reproduction noted a cult value to art.  Even early reactions to film, he notes, treat it as a cult object.

 It is instructive to note how their desire to class the film among the “arts” forces these theoreticians to read ritual elements into it

It is natural that the same tendency could be coming into play here. The algorithm is a non-human entity trespassing into the sacred space of art. Just as graven images are deemed sacrilegious in Abrahamic religion, it's not unnatural that the transgression of automated generation into art -- a realm which has retained such spiritual trappings as "genius", "revelation", and "calling" even as the rest of the world has tended toward secularism, particularly since the invention of photography -- would be regarded in the same way. Indeed, one finds among so-called outsider artists the common motivation of messages from God, and debates over what counts as "art" have continued even when hairs are split over works of human creation.

The outsider artist, in fact, may be the figure that makes me most certain that art will continue to survive. Is it possible that some people may be dissuaded from learning to paint due to the knowledge that AI can create decent looking images? Certainly, but at the same time, throughout human history, there has been a tendency for man to create art for his own gratification, even where there was no financial incentive. The first cave drawings may have been motivated by a spiritual impetus, but also must simply have been a means of expression and a rewarding pastime. As the cult significance of art has declined and artists have knowingly nurtured the authenticity of art as a source of its monetary value (Deschamps' Ready-Mades being the obvious apex of this), art as a pastime, divorced from financial incentive, has continued to survive and inspire, and doubtless will continue to do so.