AI Art IS Art

What, exactly, is art? 

According to Oxford Languages, the online dictionary used by Google, art is ‘the expression or application of human creative skill and imagination, typically in a visual form such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power.’ Some might say that, because of this definition, AI generated art cannot be art. AI, after all, is not human, and art requires ‘the expression of human creative skill and imagination.’ You don’t put pen to paper when you make an AI do the work for you, it’s true. BUT, nowhere in the definition of art does it say that art requires a human to do the creation of said art. Just that human expression and human imagination must be involved.

 A well created AI generated piece of art has every bit of imagination of one that took years to create, if not the same amount of effort or skill. But the definition of art doesn’t say anything about artistic skill- it doesn’t say anything about being good and sculpting, painting, or sketching. It says ‘creative skill.’ AI art doesn’t hinder creativity- it allows it to flourish. Not everyone has the ability to create the beautiful works of art of artists like Picasso, Van Gogh, Da Vinci, or Rembrandt, but many do have the creativity and the imagination of these art masters, just not the means. I myself, though I never have and never intend to use AI to create art for me, struggle creating my own. If you’ve seen my previous columns, many of the graphics on them were made using elements from Canva, with a few being drawn by me and one by a friend. With Canva, it’s sort of like scrapbooking, taking elements from other artists and mushing them together to create my own. Sound familiar? Generative AI does something similar, and in fact, a well created Gen AI image does it better. 

I’ve said ‘well created’ a few times now. You might be thinking, ‘Nico, what exactly does that mean?’ This is what a badly created Gen AI art prompt looks like: “Person sitting under a tree.” There’s no creativity, no imagination, no originality in a prompt like this. But in a well created prompt, it’ll look something like this: “Tall, young boy with freckles reading a red book under a reddish willow tree with curving branches as pale yellow, dull orange, and vivid red leaves fall and lie around the base of the tree. The tree casts a twisting shadow that merges with the ground.” These complicated, in depth prompts often don’t work and require much workshopping to get exactly what the prompter wants. Is this not part of the artistic process? The more specific the prompt, the more the Gen AI’s art is really just the manifestation of the prompter’s imagination. Is that not art?

AI Art ISN’T Art

Art isn’t important because it looks pretty or because it’s a masterwork in technical skill. What makes art important is the story it tells—how the artist expressed themselves through a different medium in order to convey meaning to the audience. 

Like Nico said, art is human expression. His beliefs on how AI still requires imagination aren’t entirely incorrect—generative artificial intelligence still requires some level of human input—but his flaw is that he never accounts for the process. Real art comes from human struggle, from failing to solve a problem—both technical, like learning how to paint, or personal, like identity—before finally succeeding. Art must come from a human’s hand in order to be real, because that is the only way that an artist can truly represent themselves and what they see. Even with significant human direction, AI can only ever mimic human expression, because it does not go through struggle when it’s following a prompt. AI will never know the beauty of what the human eye sees, so instruction to create a “twirling ballerina painting in an impressionist style” will never be able to live up to Degas’ ballerinas. 

Nico also argues that the more complex a human input is, the closer AI art is to becoming “real” art. He argues that there is human struggle in creating AI art, because someone might have to fine-tune their instruction multiple times before getting the result that they want.  But art is only art when it’s novel. Generative AI is a sum of everything it’s been fed through its machine learning algorithm, which means its perception of art is reliant on every photo it’s ever seen. Unlike a human artist—who can be inspired by a slew of past creations and still create new, unique work—AI creates art based off of what it knows. There is nothing new or different about what AI makes; it’s just combining human art pieces to make something that matches a prompt. 

Consider collaging, or Nico’s example, scrapbooking. Although it takes previously existing artworks—photos, drawings, etc.—it still requires a novel hand in creation. Here, the story and the struggle in the art isn’t about the creation of initial works; it’s about how the human chooses to put each piece together. The artist must first choose which images to include in their final image, and then they must choose how to combine images together. Humans have no part in the process of selecting and combining works once they input a command into ChatGPT, which is what makes AI’s process of creating “art” distinctly different from a human artist’s. 

Art is art because it is wholly human. Taking human expression and turning it into a machine-made chaos can never be real art, because it lacks the struggle and the story of the artist who created it.