Crossed Wires: AI — yeah, but is it art?

Daily Maverick

18-05-2025

Can AI gain sufficient knowledge from ingesting uncountable examples of great art to synthesise them into something novel and moving and, well, artistic?
I don't use AI to write, either as a columnist or in any other part of my writing life. But, as a frequent and enthusiastic user of AI for many other matters, I have, occasionally and experimentally, asked it to write on a subject after I have taken days, hours or even weeks to research and write a column, paper or blog on the same subject. I have been curious to see how the AI compares.
Unsurprisingly, the AI often does a sterling job — in all areas, including research, writing, narrative journey, tone and voice. In some cases, I suspect the AI does a better job than I, but I always submit my own prose to my paymasters. To my mind, it would be unseemly, undignified (even emasculating) and somehow malfeasant to submit the AI's work as my own, even with accreditation. Besides, I enjoy writing, and handing the task over to AI would diminish the pure pleasure of creating something from nothing.
There's the rub, of course. Writing is a creative activity — one that we humans alone have invented and practised. The best writers among us have aspired to enter the sacred edifice of Art with a capital A. Of all the things our species has created, it seems that Art is the most mystical — the apogee of our search for expression and meaning.
It is here that the AI/Art debate is most fiery. Can AI ever join the club of Artists? Or is there reason to believe the sign on the clubhouse door says 'Only humans welcome here'?
My original interest in AI dates back decades, to when I was a student and an aspiring jazz sax player. I wanted to play like Charlie Parker (the ultimate stretch target), but I couldn't understand why he chose the notes he did. The notes I chose to play were, at best, dull and felt somewhat embarrassing in comparison. I decided to try to use AI to uncover his secrets (and thereby improve my playing).
I ended up publishing an academic paper on the subject, and I never got to play like Charlie Parker. Even if I had uncovered why those notes were chosen by him, there remained the unanswered and critical question of how he managed to invent his solos in real time — an entirely different question. Incidentally, Parker's solos counter the narrative that AI is 'too fast' to be truly creative, or that great art needs time. Clearly, not always.
My experience goes to the heart of the AI/Art question: can AI gain sufficient knowledge from ingesting uncountable examples of great art to synthesise them into something novel and moving and, well, artistic?
Consider the following example. You are driving in a car at night and a piece of music is played on the radio — let's say a violin concerto. It moves you to tears. You pull over to the side of the road, overcome, and listen in awe to the end of the piece. Later, the host tells you that the piece was entirely AI-generated.
Before we take this thought experiment further, let me counter objections that such a thing could never happen, that AI could never do that. This is simply not true. I have heard AI-composed music (and read AI prose) that is undoubtedly moving. Those who claim that AI art has 'no soul' are projecting something they wish to be true. Without a definition of 'soul' that all (or even most) people can agree on, it's a meaningless statement. If your definition of art includes elements such as novelty and surprise and the ability to move people, then there are already plenty of examples made by AI — and they improve and multiply daily.
Intention
Back to the question posed by the violin concerto that had you sitting in tears on the roadside: Was it Art that the AI produced? In terms of the effect on the listener (or viewer or reader in other thought experiments), one could certainly argue yes. But what about intent? Surely the intention of the artist has to be a component of the entire package. This is where it gets a little hazy.
For instance, one could argue that someone was at the start line 'prompting' the AI, and therefore that human is entitled to accreditation as the artist. But this fails an important acid test because, as someone who occasionally uses AI to generate images, I know that the prompter can never predict exactly what the AI will produce. At best, we give it a nudge and wait to see what comes out, perhaps adding a few more nudges before settling on what we consider to be the 'best' output. That's hardly akin to Vermeer sitting at the easel, paintbrush in hand. With AI, the human agency is reduced to little more than an optimistic switch operator.
Worse, there are AI systems that require no human prompting at all. Just an objective like 'Make 3,000 renderings of great art — you are free to choose your own style and influences.' Certainly, among that pile, there will be a rendering that someone is likely to experience as Art.
There is a trope that has arisen in the reams of debate about this subject: that AI is simply a new and powerful 'tool' for artists to use, akin to any other tool used by artists historically. I don't buy that. A paintbrush is a tool to manifest what the artist imagines onto the canvas. Ditto the typewriter or word processor. AI is not such a tool. It is something else, an entirely independent creator of its own output. And yet, it is not an artist — not by any interpretation. It does not have a visceral need to create as we humans do.
So, finally, do we, the beholders, really need to be able to tell the difference? Do we need to know or understand the provenance of what we see or hear? Not really. If it moves you and pleases you, be grateful. We need more of that.
Do humans still need to create art? Yes. Because it is in our nature. It is who we are. It is what has always marked us out as a species. And, as for this new species now spilling into our lives, well, okay. Nice to meet you. Too bad we can't discuss art over a bottle of wine at the local tavern and then go home, take off our clothes and fall into bed. DM
Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg, a partner at Bridge Capital and a columnist-at-large at Daily Maverick. His new book, 'It's Mine: How the Crypto Industry is Redefining Ownership', is published by Maverick451 here.