The Value of AI Art
AI art programs such as NovelAI or Stable Diffusion have recently exploded in both popularity and technological fidelity. Minus the several visual quirks that come from generating a picture entirely from words, AI art is surprisingly good looking. As an artist myself, I believe that AI art opens up a lot of great possibilities for the future of visual art, but I also believe that the misuse of AI art is incredibly hurtful to the art community as a whole.
When I was researching the applications of this new technology, I came across a person excitedly sharing a graphic novel they had created using purely AI generation. This is not an insult to the author of the graphic novel, as they merely took advantage of an opportunity to enhance their work, but I was unnerved by the fact that such a thing was even possible, and even more scared of the fact that it actually looked very well made.
The most commonly discussed question surrounding AI art is whether or not it is easier than traditional art. In my opinion, it’s easier to understand the difference between AI and traditional art by looking at the skillsets required to execute each one effectively. Going by this criteria, AI art requires very few traditional art skills to execute effectively. Basic things like dexterity and hand eye coordination are obvious examples, but deeper fundamentals like composition and an understanding of light sources are not required to create a decent looking AI piece. AI art requires different skills like an understanding of the chosen program, and the ability to choose good prompts.
There is also the difference in time spent on a traditional piece versus an AI one, and while time spent is not always a determining factor of quality, the accelerated pace at which AI pieces can be made comes at the cost of trivializing many of the core parts of the artistic process.
Most of all, AI art rewards the artist’s creativity much more than their technical skill. The thing that makes art so alluring to me is not just creating inventive ideas for a drawing, but figuring out how to properly execute it. Works of art should be valued moreso for the effort put behind their creation rather than the ideas it represents. Subtext should never be the main draw of an art piece, since the ability to assign meaning to things requires none of the technical ability that well made art requires.
Even though the process behind making AI art is faster than traditional art, casually experimenting with programs to work on personal projects is a harmless practice. What makes the graphic novel, and any other AI generated piece of media, worrying is the fact that they can and have already been sold for profit.
The intent of this article is not to label any technological advancement as harmful for art. I primarily do digital art, which is objectively more convenient and easier to use than most physical mediums. The difference between digital art and AI art is that while they both make the process of art easier, digital art still requires an understanding of art fundamentals in a way that AI art simply doesn’t.
Although many of the opportunities afforded to artists by AI art are beneficial, the potential for misuse means it is more often used for harm rather than good. As the products of AI generated art become more and more sophisticated, it’s important to recognize the utility of these programs, but not allow them to be used in ways that hurt other artists.
By Christopher Eckl