The Ethics of AI Generated Art
Chances are you’ve already seen the headline (or some variation thereof):
“AI won an art contest, and artists are furious”
Here’s what happened: A Colorado man entered an art competition at the Colorado State Fair Fine Arts Competition in the category of “digital arts/digitally-manipulated photography”. The problem, however, was that he produced the image using Midjourney, an online AI program that produces images based on user text input. He entered the piece using the name “Jason M. Allen via Midjourney”, thus disclosing the use of the AI and meeting all competition rules. And he won! The judges were not initially aware that AI was used, yet later admitted that they still would have awarded Allen the prize even if they had. As the headline above attests, many artists the world over were not pleased with this outcome. In fact, many were livid.
This story is an apt touchstone for the emerging debates on the ethics of AI-generated images. While artists have good reason to be angry with the outcome of that content, I am not sure it is for the reasons they think. Like many people engaged in this debate, a lot of anger and energy is being misdirected, missing the heart of the issue. It’s important to note from the start that this will not be (and cannot be) even close to a comprehensive take on this topic. To be sure, volumes will be written in due course. Further, I am not an expert in art, ethics, copyright law, or any of the other fields. And finally, I have not landed on a firm position on all aspects of this debate.
However, as someone who produces and sells digital art, creating art for both myself and for paying clients- not to mention as a professional writer in a field where AI is beginning to emerge as well- I do have a stake in the game. Further, I have spent the last few months (bed-ridden due to prolonged illness) obsessively exploring these AI programs, especially Midjourney, and even selling some of the works I developed.
So with all that said, let’s dive into some of the points that have been missing in much of the conversation.
The Art and The Artist
The question of “What is art?” has been raging for centuries and won’t be resolved any time soon- and most certainly not here. So for the purposes of this piece, let’s reference a very basic definition while acknowledging its limitations. A common definition of art says that it is: “a diverse range of human activity, and resulting product, that involves creative or imaginative talent expressive of technical proficiency, beauty, emotional power, or conceptual ideas” (from Wikipedia)
Based upon this definition, AI-generated images are, indeed, art at its most basic. They meet all the criteria, if in unorthodox ways. If you look at much of the work being produced, you would be hard-pressed to argue that it is art in the most open and broad sense of the word.
Instead, I think the question that is more at the heart of the issue is this: “Who is the artist?” If I enter the text prompt with all its potentially unique perimeters, does that make me the artist?
In my opinion, yes and no, but perhaps mostly no. Let me explain.
The AI (aka its programmers) is bringing technical proficiency, drawing on the creative and imaginative expression of countless other artists, plus learning from all the other users. I certainly participate and contribute to the process, but I am a relatively minor player in it. At this stage (especially if I do not alter or edit the image through my own digital manipulation), I think of myself less as the artist and more as the art director. However, an argument can be made either way, especially as the user’s role becomes more involved as the technology develops.
Some will go so far as to say there is no artist at all since a machine is creating these images. This certainly brings more complexity to the questions, for again, the proficiency we associate with the artist is now “mechanical”, but I am not so sure this is entirely new either. Early photographers were similarly denied the category of an artist by some because, it was said, a machine was simply duplicating reality. In other words, critics did not see or value enough human proficiency in the creation of the photograph to deem it art or the photographer an artist. Later, even after photographers were firmly ensconced as artists in the collective view, digital photography (and digital editing) created yet more debate.
From Michelangelo to Andy Warhol, the integration of technology has consistently shaken the established assumptions and expectations of art and artists. In time, we generally adapt. That does not mean current concerns about AI are unimportant, only that they are not unprecedented in the broader sense. And that very knowledge should give us pause and invite some tempered consideration before rushing to judgment.
Whose Art Is It Anyway?
Another major concern- and one with perhaps the most merit- is where the AI gets all the data with which it learns. Since it is drawing from all available images on the net (and probably beyond), that means it has access to the works of all the greatest artists, alive and dead. More than that, it has access to all images by all artists that are in available digital form. Not only does it learn from these images but the user can direct it to draw from the very styles of specific artists. At what point does the line between “inspired by” and “stylistically stolen” get crossed?
For all practical purposes, this question is most likely going to be worked out in the legal arena, as artists (and their estates) push back against unlimited use of their work in this way. While I will be watching those battles closely, they won’t really be resolved the underlying questions fully (or immediately enough), so I will leave those to the lawyers. After all, even if it could be shown that the technology crosses a line, is it the responsibility of the technology to limit the reach? Or is it the user’s to use it responsibly and legally? Not an easy question.
Instead, however, we must acknowledge that all art is derivative- at least on some level and to various degrees. Whether consciously or not, we are all working from the collective data we accumulate. We don’t shield art students from the works of great and diverse artists for fear they will “steal their style”, but rather rely on it to enrich and inspire them. In the end, aspects of those artists will shape their work. Further, the unconscious influence of visual learning on every individual is impossible to measure. Our brains are constantly integrating information at unfathomable rates.
To be clear, I am not saying that this justifies artistic copying and/or claiming other artist works. However, technology like Midjourney doesn’t actually use or reuse any of the images it draws from. Instead, it learns in a way more closely related to our own assimilation of images. (To understand more about how the technology works, take some time to watch this video:
Again, this does not resolve the question but debunks the claims that the AI is simply “copying” work. It does, however, make “the line” that much more difficult to discern.
Save the Starving Artists
In the end, most of the pushback I have heard has come from artists whose core concern is that they will be robbed of their livelihood by a technology that stole from their proficiency and productivity, allowing the AI users to create more content of seemingly higher proficiency in a fraction of the time, with little talent, training, or vocational passion. This is a concern that I fully understand. As I said earlier, as a digital artist (albeit in a very basic and minimal sense) and as a professional writer, the challenge of making a living with my work seems impossible at the best of times, let alone when computers are learning to do it too.
However, I am not sure the assessment is entirely fair, in part because these problems are not entirely new. Every technological development challenges the pre-existing order. Painters were threatened by photographers. Radio threatened newsprint. Video killed the radio star! And while this does not negate the merit of the concerns, I am unsure that an alarmist response or an absolutist posture of opposition are either helpful or fair.
Instead, one of the biggest problems I see around the question of the artist’s livelihood is the failure to critique the devastating impact and problematic enmeshment of capitalism and art. So much of the discussion has centered around how artists make their living, as well as how (explicitly or not) many associate the quality (aka value) of art relative to its earning power. That is a capitalism problem more than an art problem. An artist’s ability to create work and the value of that work should not be tied so significantly to consumerism. After all, do most of us believe the commercial success of the 50 Shades series is based on its inherent literary quality? Not so much.
I’m not being naive. I know we can’t simply flip a switch and separate art from capitalism. Nor am I denying that monetary value is an important part of art insofar as it can provide fair(ish) compensation for labour. Rather, what I am suggesting is that we must be careful not to put the bulk of our attention critiquing this technological development (and those who utilize it) and so little on how capitalism has failed art (and the artist) time and again.
A New Renaissance?
What is being overlooked by so many in the emergence of this technology is the marked increase in interest in visual arts. People who would otherwise give little thought to appreciating (letting alone creating) art are suddenly excited, curious, daring, and engaged. Even something as simple as a Dungeons & Dragons group utilizing the technology to bring creativity and vibrancy to their game is something exciting. When the broader culture begins to engage with art, all art and artists win.
Further, this tool is being used a great deal by established and emerging artists. I have seen countless images of paintings produced by artists who got inspiration from an AI image they prompted. I have seen authors use AI images as visual prompts to inspire their creative writing. I have seen digital artists utilize bits and pieces of their AI images, incorporating them into their other work. Artists are already deeply benefiting from this development and I only see that increasing as we figure it out.
In the end, we need to celebrate that this technology is opening up millions of people to the beauty and complexity of visual arts. People are learning about composition, perspective, lighting, and the impossibility of hands and eyes (if you know, you know).
So, as we engage with the legitimate and complicated challenges of this new emerging technology- dare I say, art form- let’s choose to embrace a cautious optimism and resist the impulse to react with fear and anger. For if history has taught us anything it is that everyone loses when we seek to serve as gatekeepers to art.
Jamie Arpin-Ricci is a bisexual author, activist, and the Co-Director of Peace & Justice Initiatives. You can discover more about his work at his website: www.jamiearpinricci.com