Can AI paint electric sheep?: Some thoughts on art generated by artificial intelligence
A big part of what makes art interesting is the thought behind it, and that shouldn't be controversial
Have you ever watched a movie and thought, 'Why did the actor pause after that line?' 'Why did they look away from the camera—or towards it?' 'Should I read into it?' 'I wonder what it means?' 'Is it just a coincidence?'
If you're watching an actor of a certain caliber, you have an understanding and trust that, whether or not what the actor is doing works for you or for the moment, that they're doing it intentionally — that there's thought behind it.
With the rising popularity of AI art applications (many of which are well known for stealing work from artists through scraping the internet to feed the AI), there's also been a rising conversation about what the value of art is, and just as importantly, what the role of the artist is. While some are dismissing artists in this conversation, comparing their work to the work of other craftspeople in industries that have become mechanized, to do so would be to ignore what art is.
Art is not simply its final product. It's a living thing, dependent on an understanding between a creator and the reader, one that promises the reader that the art is worth their time. So what does that mean in terms of the use of artificial intelligence in art?
Does Artificial Intelligence have a place in art?
I believe that there is space for AI in art— some of my favorite recent works of art have artificial intelligence as a primary element of the piece.
Consider The Substitute by Alexandra Daisy Ginsberg, a life-sized projection of the extinct northern white rhino, brought back to life through research from AI lab DeeMind. With the use of this data, the rhinoceros learns from its surroundings, changing its behavior, and as time goes on, its presentation also becoming more pixelated, details fading to match data of the shrinking numbers of the species, until it, like the northern white rhino, disappears entirely.
Now what sets the Substitute apart from the AI art programs that have been taking social media by storm? Well, for one, The Substitute is an example of AI art with with purpose, and the AI does something that an artist cannot, it replicates a type of life in a way that speaks to what the artist is trying to represent. The Substitute uses AI to say something, making it markedly different from the AI art that is at the center of the conversations artists are having right now.
The AI art created by programs like MidJourney and Lensa AI, take aggregate information (often stolen art) and generate 'new' images from inputted keywords. You could type in 'sad monkey,' and the program will spit out an image of a sad monkey. Or dress it up a little and type up 'sad monkey wearing a hat in the rain,' and so on. For your keywords, you are rewarded instant images from the algorithm. But are these images art? Well, sure, in a way. But is the art compelling? Is it useful?
I don't think so.
There is a difference between 'instant art' and what I would consider real art — and it's not one of skill or of gatekeeping or snobbishness; it's a difference in intention. Instant art, art without consideration, without thought, lacks the backing that makes art interesting.
Meaning is made through the process
The difference between someone plugging in words to create instant art and an artist is that an artist is forced by the process of making to consider what is going into their work. They must continuously making decisions about what variables get plugged in and changed and adjusted. Sometimes this means choosing a medium. Sometimes this means deciding how much white space is left on a page. Sometimes that means figuring out specifically what data (hopefully not stolen) is plugged into an AI database.The point is, that it's part of the artistic process to make decisions.
As interesting as AI art can get, like regular art, it only holds value with an artist making real process and concept-based choices behind it. Thought is what separates Félix González-Torres's 'Untitled' (Portrait of Ross in LA), a memorial for González-Torres partner who died of AIDS from a pile of candy, what Cameron Rowland's Lynch Law in America, a commentary on the act of calling the police as a "request that someone be injured, arrested, incarcerated, or killed." from a regular emergency call box.
What color emergency call box should be used? Should it be altered? At what height should the box be hung? Should the museum replenish the pieces of candy after they're taken away from the sculpture? These are only a few of the questions that an artist must face and decide, and their choices inform how the viewer understands the art and the artist.
But the viewer is only going to do that deeper work, that deeper thinking and interaction with the art because they know that the artist has put the value in, has (at least promised) to make it worth their time. Without those decisions, the thought, the consideration, no matter how "good" instant AI art can get, it will never hold the value that art made by people holds.
While we can definitely connect to and develop relationships with technology in many ways (shoutout to the Mars rovers), art is interesting and compelling because its made by people who are spending their time and effort to create an effect. And what makes that interesting can't be replaced by a program that wipes away the effort, the decision making, and the thought. At the end of the day, because there is no time put into making instant art, instant art simply isn't worth your time.
CBLDF: The book ban movement is aiming to demonize the 'graphic' in 'graphic novel'.