Over the course of 2022, few things outside of electoral politics have been quite as divisive on social media as AI-generated art. This technology, though it almost seems to have sprung out of nowhere, has been in development for years as part of the broadening scope and capability of “deep learning” algorithms. Although it is still very imperfect in some respects, the ability to create images from simple text prompts has shaken the foundations of the world of art, leading to heated debate over its validity as “art” and great consternation regarding the fates of artists. As the technology becomes more refined it will no doubt alter our culture in irreversible ways, forcing us to come to terms with it whether we like it or not.
When the image-generating software known as DALL-E Mini (now called Craiyon) went viral earlier this year it was in part because of its capabilities but also because of its limitations. You could type in a prompt like “The Muppets in Saving Private Ryan” and get results that indeed seemed to place Jim Henson’s beloved puppets into WWII uniforms with Thompson submachine guns in their hands, but usually with certain visual mutations that made the images easy to identify as algorithmically generated. The humor that came from these was largely based on this identifiability. One set of images shared on social media was based on the prompt “Gordon Ramsay shouting at a child for putting a potato in a urinal.” If one merely saw a photograph of this happening in real life, it would not be nearly so entertaining as seeing a computer trying to figure out what the hell the scene is supposed to look like and returning bizarre images of a vaguely Gordon Ramsay-like shape pulling vaguely vegetable-like shapes out of a vaguely toilet-like shape.
Things took a slightly darker turn when more refined pieces of software were released a few months later, including DALL-E 2, Stable Diffusion, and Midjourney. These new, more powerful AIs are capable of generating images less distinguishable from real photographs or artworks. Though they still struggle with certain details (for instance, hands with the right number and arrangement of fingers), the technology has been rapidly improving and is already being adopted by a number of businesses and websites. The Economist used Midjourney to create the cover for their June edition, and an image created with the same software won first prize (much to the chagrin of the other entrants) in an art competition at the Colorado State Fair in August. DeviantArt, one of the internet’s oldest and most popular sites for artists, openly embraced AI art, recently adding a text-to-image generator to their site based on the Stable Diffusion software and automatically feeding all of the hundreds of millions of artworks uploaded to the site by users into its database.
These events immediately caused grave concerns for traditional artists and their supporters (“traditional” here referring to those who use both physical and digital media, but do not use any sort of AI in their work). The thought that AI-generated art might copy one’s work seems an affront to those who have dedicated their lives to developing their skills and style even more so than if someone merely traced or reuploaded their art and claimed it as their own. Concerns have only grown as some AI art enthusiasts have shared the prompts they enter to generate their images, revealing that they use tags such as “trending on ArtStation” or even a specific popular artist’s name. In some instances the software has even attempted to copy the artist’s signature onto a work that they did not in fact create. Artists who livestream themselves spending hours creating digital paintings have been subject to their work being screencapped and fed into an AI before they are even done with it. One illustrator on Twitch had their work taken and put through an AI by a viewer who then claimed that their work was a copy of the AI-generated image, almost certainly causing great distress for the artist whose reputation was at stake.
The ethical questions raised by this new technology are numerous, as it poses a direct threat to traditional artists’ livelihoods. It may not be such a big deal to have an AI generate a Picasso painting of a scene from Super Smash Bros., but to accurately imitate the style of someone who is currently trying to make a living off of creating art is to risk robbing them of their career. As president of the Graphic Artists Guild Liz DiFiore states in an article on ArtNet, “infringement is probably the biggest reason why illustrators are seeing the value of their work fall.” The art world is competitive as it is, and now that artists have to compete with algorithms that can copy their hours, days, even years of hard work in a matter of seconds, the viability of creating art as a human being has taken a devastating blow.
In a broader sense, as the technology behind AI art continues to improve, it will without doubt bring irreversible changes to the way creative minds fit into society in general. The developers of DALL-E 2 have predicted as much, stating that their software “could displace jobs of designers, photographers, models, editors, and artists.” It’s likely that there will be a day in the not-so-distant future when, instead of having a full team of graphic designers, a corporation will be able to subscribe to an AI art service that can generate marketing materials instantly on demand. AI-generated literature will eventually saturate the market and make it increasingly difficult for actual writers to gain an audience. Coupled with procedural generation and speech synthesis technology, it’s likely that films, TV shows, and even video games will all be made by an AI someday, since this will be much cheaper than paying a full team of writers, artists, and actors. And consumers will inevitably have access to the technology as well, enabling them to create images to their liking without commissioning an artist and to edit their photos without paying someone who is skilled in Photoshop.
If one only considers the economic side of things and the sheer convenience that the technology will provide, AI art with all of its potential comes across as a godsend. The ability to generate images both mundane and fantastical essentially at the push of a button will revolutionize media forever, will save millions of dollars, and will drive up profits exponentially. As is typical with capitalism, this is all that really matters in the grand scheme of things. The lives and livelihoods of traditional artists and writers and the importance of genuine human creativity will hardly even be an afterthought—but this is no surprise to those who are familiar with the history of capitalism and its propensity to simply toss aside anyone and anything that is not profitable.
It would be easy here to continue elaborating on the potential evils that AI art brings in its wake. It is, after all, only the latest instance of capitalism’s parasitic nature, its desire to latch onto and profit off of every aspect of the human experience it possibly can. And while such an analysis would be worthwhile, there is something else that needs to be considered which doesn’t seem to be present in most conversations about AI art. Whether the new technology is ethical, whether it should be used, and whether the things created with it even qualify as “art” are not the questions we need to be asking. What we need to consider is how this new technology will forever alter our culture, and we can look back at similar technological developments in the past for some guidance.
Walter Benjamin opens his essay “The Work of Art in the Age of Mechanical Reproduction” by stating that art has always been reproducible in one way or another; copies of certain works have often been made by students learning from their masters or by those looking to sell duplicates of a work. The advent of printing, photography, and film made the reproduction of art easier, but this was not all the new technologies were capable of doing. As they developed and became a normal feature of society, they reached a point where they “captured a place of [their] own among the artistic processes.” What these technologies gave us was not only the ability to reproduce art mechanically but to produce it mechanically as well.
What is worth noting here in Benjamin’s essay is his statement that “the nineteenth-century dispute as to the artistic value of painting versus photography today seems devious and confused.” He explains that the full implications of the new technology simply could not be realized at the time, though the uproar over it was suggestive of how dramatically things had forever changed. Amidst the debate between cultural critics, “the primary question—whether the very invention of photography had not transformed the entire nature of art—was not raised.”
The fact of the matter is that there was a day when it was thought laughable that a photograph would ever end up in an art gallery. Early critics insisted that film was nothing more than a mindless form of entertainment, and today there are many who still consider video games to be a useless, childish pastime in spite of the artistry that goes into creating them and their capacity for narrative immersion. Yet today it is not surprising to find photographs in museums of fine art, and the cultural impact of movies and video games hardly seems unwarranted (even if some of the most popular instances of these forms of media have become repetitive, easily marketable rubbish). And this recognition of how our cultural perception is altered by new technological developments is precisely what is missing from most current debates on the ethics and validity of AI art.
Yes, machine-made art has terrifying potential to devalue human creativity and rob artists of their livelihoods. But the question of whether or not it counts as “art” is not so important as the question of how we, as a society, are going to deal with its existence. No matter how hard traditional artists protest, no matter how vehemently we might resist the new technology, it’s not going anywhere. Something so profitable will never simply vanish. And while it is difficult to say exactly how AI art will alter our culture in the future, it seems undeniable that someday soon it will become as commonplace as all of the aforementioned technologies. Rather than arguing over its virtues, what traditional artists and their supporters need to do is figure out how we are going to live with it. We need to recognize that its ascendancy—like that of text messages over telegrams—is virtually inevitable at this point, and we need to come up with ways for traditional forms of art to survive and compete.
Unlike certain other recent developments in digital technology—particularly Zuckerberg’s Metaverse and the NFT craze, which were both impractical from the start and have been almost universally panned—AI art has immediate practical applications and consumer appeal. As difficult as it is to accept, efforts to resist the progress of this technology will eventually be seen in retrospect the way we see those early critics of photography and film who denied their validity as forms of art and who denied the virtue of those who initially appreciated them. Just as the printer threatened the job of the scribe and the photographer threatened the job of the portrait painter, today the algorithm threatens the position of all creatives in society. So, what do we do? Do we become 21st century luddites and make attempts at sabotage, or do we embrace the changes and plunge head-first into the new world of machine-made art? Is there a middle ground we can strike where traditional art can live on beside AI art despite the former’s economic impracticality?
What must be remembered, whatever decision we make, is that the human need to create art will not die, even if technology and economic demands make it increasingly less viable to be an artist in the conventional sense. Creativity is embedded in our nature. We will always feel the drive to turn imagination into reality, to turn suffering into beauty, to turn the old into the new. As Friedrich Nietzsche once said, “art is the highest task and the true metaphysical activity of this life.” While our ability to pursue creative passions is constantly overshadowed by expectations of productivity, profitability, and practicality, those passions, which transcend the utilitarian demands of day-to-day life, are perhaps the greatest source of fulfillment and satisfaction we have access to.
AI-generated art will no doubt make life increasingly difficult for traditional artists over the coming years. There will be more instances of infringement, more instances of mimicry, an increasingly competitive and constricted field of work, fewer customers, less income, and heightened anxiety about the future. But as the economic value of art decreases, its existential value increases. As more and more of our lives become absorbed by the increasingly inhuman mechanisms of capitalism, as our jobs treat us like replaceable drones (and in fact have begun replacing us with drones in certain industries), as our interests and preferences are fed into marketing algorithms, as the things we create are taken and mutated by generative software, as we are surveilled by automated systems in hopes of mitigating any dissent, we long for liberation. And what can be more liberating than coming in contact with other human beings through the things they have created with their own hands, with their authentic thoughts and feelings as the raw material? Art, literature, music, poetry—these are what make life worth living, and they are our best hope for salvation as the shackles of a profit-driven society grow tight around our ankles.
-Jordan
Header image features Van Gogh’s Starry Night flanked by several AI-generated imitations.
Human, Being 