Should we redefine art now that, increasingly, AI is being used to create it?
Words John McDonald
I once thought it would be interesting to put together an anthology called “Art is”. Every book seemed to offer a new definition, leading me to imagine what a huge volume of such definitions must exist, many of them mutually exclusive. Does AI abolish these countless definitions, accumulated over thousands of years? Of course not, but it adds a new layer of complexity.
It’s not just art that is rendered problematic by the AI revolution, it’s virtually every walk of life, with the possible exception of sport. We need to assume that AI’s fast-learning capacity means it will be able to write fiction, poetry and journalism, compose music, and design everything – from frocks to buildings to urban transit systems – with a level of competence that far outstrips the average human being.
In art, the rise of AI-generated work coincides with a widespread feeling of creative exhaustion, and the triumph of commerce over aesthetics. This is a topic worthy of a lengthy discussion, but suffice to say there is a global sense of predictability, as a small group of powerful dealers dictate to public museums who or what they should be buying. Meanwhile, the market is bloated by newly coined billionaires buying almost indiscriminately in search of status and investment opportunities.
This is also a time when many museums are struggling to attract audiences as new generations, accustomed to getting their experiences virtually, hardly recognise the difference between viewing works on a gallery wall or on a screen. Many would argue that the screen is more convenient and doesn’t incur hefty entrance charges. There’s also a problem with shrinking attention spans. For a huge number of people today, museums are simply boring.
What AI offers is not merely a new form of artistic expression, as pioneered by figures such as Beeple, who creates and posts a satirical image every day; or Refik Anadol, who makes large, ever-changing abstractions from unthinkable volumes of data; or even Ai-Da, the humanoid robot, who paints with mechanical arms. AI offers the possibility of making your own art in an unlimited range of styles by absorbing and manipulating the data of art history. The machine does the work but human actors will claim the credit, and potentially the profits.
When Ai-Da’s portrait of AI pioneer Alan Turing sold for $1.6 million at a Sotheby’s auction in November 2024, it was widely considered a turning point in the acceptance of the technology. This might be true, but if so it’s a dismal reflection on the cultural decadence of our age.
It might be the case that more and more art is being created by AI, but most of this work is – to venture a quaint, romantic term – soulless. The cerebral, amoral nature of the medium has proven to be highly efficient at creating images, but if these dazzling performances feel lifeless and hollow it’s because the machine is not yet capable of making huge mistakes, doing irrational and temperamental things, or changing its mind in mid-stream. As it currently exists, AI could never have painted a hybrid picture such as Picasso’s Les Demoiselles d’Avignon (1907). Had it existed in the 19th century, it could not have come up with anything as daring and rule-breaking as Manet’s Le Déjeuner sur l’herbe (1863).
AI today acts as an enormous compendium of images and styles, thrown together in ways that might seem bizarre – not because they are irrational, but because they are too rational and fact-based. This is “art”, inasmuch as anything can be nominated as such. History and philosophy suggest we’ll never be able to agree on a more precise or exclusive definition, let alone one that excludes the use of new technology. As technology changes our lives so, too, does it change our ideas about art, although not necessarily for the better.
My dystopian scenario is a Terminator one, where pockets of painters and sculptors hold out against the AI revolution, making work for small groups of connoisseurs and hobbyists who doggedly insist that a work of art is more than an image. Those rebels form a kind of resistance, awaiting the day when the sheer weight of instant entertainment provided by AI begins to make people feel glutted, confused and exhausted.
From out of the shadows, the Slow Art movement arises.
Words Alison Kubler
As I write this, I’m fighting a yawn because I took my daughter to see composer Max Richter in concert last night. It was, just as I expected, extraordinary. I spontaneously began to cry; in the dark, I glanced at Olympia and saw tears rolling down her face, too. I gently touched her arm. We both wept silently. I noticed people around us also wiping away tears and sniffling quietly. It was the epitome of collective effervescence, a term coined by a French sociologist to describe a shared experience of connection and a synchronicity of euphoria. I thought about how lucky I felt and how grateful I was to receive the gift of music from the musicians on stage.
Unable to sleep, I devoured everything I could read about Richter’s oeuvre and came across his recent comments regarding the threat of AI to musicians.
Richter’s observations, presented to an audience of British MPs, were nuanced and tied directly to the training of AI using copyrighted music, which infringes on artists’ rights – a concern shared by many creators, including visual artists, writers, and designers. The composer delivered a passionate speech defending artists’ rights while clarifying that he is not “anti-AI”. This seems to be an important distinction. He explained, “There are many areas of life, especially in scientific and technical fields, where it’s obvious there are huge benefits to using this technology ... However, the situation is more complex when it comes to
creative work. We need to approach it thoughtfully and consider the societal impacts that its use in creative fields could have.”
These are vital discussions for the era we inhabit. This marks a turning point for creatives, who are often not adequately compensated for their work.
How do we teach AI to respect intellectual property? As a young art historian, I encountered theorist Walter Benjamin’s 1935 essay, The Work of Art in the Age of Mechanical Reproduction, which argued that the use of machines (AI could fit here) devalues the “aura” of an artwork, that intangible “thing” that connotes authenticity and gives an artwork its originality. Revisiting it now, I am struck by how prescient it was and how the fear of the new persists.
The invention of photography was heralded, much like AI, with similar disbelief. Naturally, photography did not prove to be “the death of painting”, as many art historians foresaw; instead, it gave birth to a new way of seeing, leading to the
emergence of Impressionism.
Artists are often early adopters of new technologies, eager to explore their applications and potential. Increasingly, artists are using AI as an artistic tool to create and produce art, much as they did with photography and, later, video and computers. They remain artists and what they create, even with AI, is still authentic.
Can AI make art? Yes, of course. Will it be “good” art? I don’t know. Good artists make lousy art, too. I don’t think AI can write music like Richter. But maybe it could? Will it be authentic? Yes ... and no. Will it have that “aura”? I don’t believe it will.
What I do know is that art is the most undefinable of things. I don’t want to exist in a creative world with parameters about what constitutes an artist, because I think everyone can be an artist. AI “artists” already exist. I’m reminded of Andy Warhol’s 1963 deadpan exhortation, “I want to be a machine, and I feel that whatever I do and do machine-like is what I wish to do. I think everybody should be a machine.” I suspect Warhol would have loved using AI as a tool. Of course, he also said, “Art is what you can get away with.”
John McDonald writes a weekly art column for The Nightly, and publishes Everything the Artworld Doesn’t Want You to Know, at jmcdartcritic.substack.com
Alison Kubler is an art historian, curator and editor-in-chief of VAULT, vaultmagazine.com
