Someday we will all be artists

How AI will change the nature of work and art

I am a co-director (with Canice Prendergast and Mary Ittelson) of the Arts and Creative Enterprise Initiative at UChicago’s B-School. The program sits at the intersection of arts and business; the aim is to bring the artistic mindset into the practice of business and management through coursework (co-taught with the art department) and events.

Here is the question I get asked most: how AI will affect art and artists?

My answer: AI will be incorporated into art, but it will not replace artists. In fact, as more and more human labor becomes automated, there will likely come a day when most if not all of us will end up doing work that right now would be classified as art, or at least artisan. Why?

Rock will never die (the Humanness Premium)

Art is part of a category of goods—experiences, really—where the humanness is central to what is being consumed. When a person buys, views, or engages with a piece of art (which can include music and film) they are not simply consuming content. They are entering into a relationship with the artist. They want to know the story, the process, the background. Just look at how many celebrity magazines, documentaries, and biopics are devoted entirely to understanding the person behind the work. If anything, the global appetite for artist origin stories, studio tours, and behind-the-scenes footage is enormous and growing. People don’t just want the output; they want the human journey that produced it.

Last month I discovered what is now my new favorite band: Geese. I very quickly became enamoured. I listened to their album front to back, then back to front, over and over again. Songs like Cobra and Husbands got stuck in my head and would not leave. Okay, so while the album seemed completely unique and original to me, maybe in some not-so-distant future an advanced AI could pull this album out of thin air. This is absolutely possible. Why do I think artists won’t become replaced by AI? Because here is what else I did almost as soon as I heard (and loved) the first track on the record: I obsessively began to research everything about the band. Their love for Midwestern-Emo, where the name Geese came from, their use of a looped Ukrainian Choir in the title track. The more that I learned about the band, the deeper I fell into the contours of the music. The grooves, samples, guitar riffs, and most importantly, lyrics, all took on a different meaning. I don’t think I’m anywhere close to alone in this experience; most people who consume art in some way report a similar symbiosis between the artist and the art.

What does this all mean for art and AI? What I think it means is that the “humanness” of the art—what Séb Krier calls the “touchy feely” dimension of a product or experience—is just as important to the consumption experience (in the case of modern/post-modern art, more so) as the actual aesthetic component. Séb makes this point nicely in a guest essay on my Substack: he argues that demand for “touchy-feely things is not an inconvenient detail”—it is a core feature of how human economies work. Even in a world with very advanced AGI, where machines do the vast majority of the work, the remaining human activity can still be well-paid. As he puts it, you cannot “mothball” humans if the consumer preference is specifically for the humanity of the human. If you model humans as utilitarian agents who only care about efficiency, then yes, they become a hindrance; but if you model them as demanding other things—the process, the taste, the conferral of status—then the “transaction cost” of a human taking 40 hours to make something instead of a robot taking 4 seconds is the value proposition.

Séb also makes a point that I think is crucial: in a post-scarcity world, provenance remains scarce. Just because an AI can produce a technically flawless reproduction of the Mona Lisa doesn’t mean the original sells for any less. People pay enormous sums to attend live concerts even though listening to the same music is much cheaper. If an AI model created an album just as good as Midnights, do we really think people would pay $6500 for a ticket to see it performed live? No, you need Taylor Swift and her entire journey as an artist for that. The being able to perfectly replicate human creativity is simply not sufficient to eliminate demand for the real, human thing.

The upshot is that as AI automates an increasing share of commodity production—of tasks where the identity of the producer is irrelevant to the consumer—demand will naturally shift toward goods and services where human identity, story, and craft matter. In a wealthier world, with more leisure time and more disposable income, demand for authenticity, for status goods, for human-made experiences, will grow. Remember, six out of ten jobs people are doing today didn’t exist in 1940—it is not unreasonable to think that post-AGI, many future jobs look like what we think of as artists today, where the “humanness” is the value.

We have been here before

Why am I fairly confident about this? Well, we’ve already lived through a major disruption in art: Photography.

Before the daguerreotype was introduced to the public in 1839, one of the main purposes of painting was to faithfully reproduce reality. For centuries, an artist’s skill was evaluated as their ability to mirror the natural world in their art; by the early nineteenth century, the Pre-Raphaelites and French Neoclassicists were able to accomplish incredible feats of realism with the paintbrush. Just look at this William Trost Richards painting below. Part of what made a great painter great was their ability to faithfully recreate the world; these recreations took years of training and could only be produced by skilled human hands.

Photography automated that task almost overnight. A machine could now do, in minutes, what had taken artists weeks or months to accomplish. The reaction was: panic. The French painter Paul Delaroche, upon seeing the first daguerreotype, apparently declared that “from today, painting is dead” (though this quote is disputed). And indeed the disruption was real. Many portrait painters did lose their livelihoods and clients switched to the less expensive photographs. By the 1850s, Parisian studios were producing thousand upon thousands of photographic portraits a year—products that only a few years before could only be made by expert painters.

Did painting die? No, quite the opposite. What happened afterwards was arguably the most explosive period of creative innovation in Western art. Art was no longer constrained to reproduce reality—a camera could do it well enough. The realm of the artist was everything that photography could not do. Impressionists began to capture the process of seeing, the experience of it. How the light plays on the haystacks at dawn, and then at dusk, as in the paintings of Monet, aimed to capture the feeling of a fleeting moment.

In one of the greatest quotes of all time, Edvard Munch wrote, “I have no fear of photography as long as it cannot be used in heaven and in hell.” Van Gogh described his artistic breakthrough as being driven by explicitly distanced himself from photographic realism.

From the Impressionists came the Post-Impressionists—Cézanne, Gauguin, Seurat—those gave way to Cubism, Expressionism, Surrealism, Abstract Expressionism, and on and on. With each movement we went further and further away from representational fidelity and closer to the place where human subjectivity and emotion live. You only need to spend a bit of time in front of Otto Dix’s paintings of World War 1—which he spent in the trenches—to be moved, to leave your own body and enter a space that you had never occupied before. But his surrealist/impressionist style only became possible after technology gained an upper hand in depicting realism.

The art world we see today is immeasurably larger, more diverse, more ambitious, and (if you include film and music) more culturally central than it was before the camera arrived. More people make a living as artists—broadly defined—than at any point in human history.

The lesson here is direct. A technology that automates the mechanical dimension of a creative practice does not eliminate the practice. It transforms it, and often expands it, by forcing practitioners to discover and develop the dimensions of their work that the machine cannot reach. The camera could reproduce appearances, but it cannot replicate the narrative of the artist. And it turned out that what people really valued all along was not the reproduction, but the expression.

What will the new art look like?

Yesterday Scott Kominers sent me a post that is one of the best clues into how art will evolve with AI. It features the work of Chema Mendez, a digital collage artist based in Spain who goes by mendezmendez. Mendez has been creating mixed digital media works since 2009, merging old paintings with modern source material to create surreal, layered compositions. If you look closely at his art, you can see the math of M.C. Escher, the surrealism of Magritte, and the digital collage aesthetics of Eduardo Recife. Mendez has been using AI in his art for years. Seeing how he uses it, it becomes clear that asking “will AI replace the artist” is the same as asking “will the camera replace the artist”: AI is simply one of many materials in his toolkit, which includes public domain paintings, 3D renders, his own sketches, and yes, photography.

The post Scott sent me is a perfect case study of AI’s role in art. Mendez was commissioned by UPenn’s Psychiatry and Neuroscience department to create a piece of art celebrating a multi-year study on pain and emotion published in Nature. The piece, titled “Relief” (below) illustrates a novel gene therapy pain treatment, which is delivered through a virus to the prefrontal cortex. Importantly, the treatment mimics the analgesic effects of opioids without actually using any. But it is the story of the artist, whose own family suffered from opioid addiction, and how he created the piece that gives it gravity and depth.

Here is the thing: Nature has a strong policy against the use of fully AI-generated imagery. So Mendez got creative. All of the images are taken strictly from the open domain. The poppy flowers—representing opioids—are from Gustav Klimt’s paintings and biodiversity illustrations from the 1800s; the human figure is a composite from two drawings and art by David Humbert de Superville and Edouard Goethals. The cortical neurons and shape of the virus are recalled through the geometric shape (an icosahedron) connecting the poppy flowers, within which the human figure appears serene like a blue sky. Every single element in the piece was deliberately chosen, researched, and assembled by Mendez to encode specific scientific meaning through artistic composition.

Creating this piece required making dozens of creative decisions—what source material to use, how to represent a virus, how to symbolize pain relief through flowers, how to layer classical drawing and scientific imagery into a coherent visual narrative. Creating the piece required knowing art history, science, as well as having the personal emotional investment to understand the importance of opioid-free pain relief. AI can surely generate an incredible image of a figure surrounded by flowers. But it could not make this piece, because this piece is inseparable from the person who made it, and the humanness which gives value to the work.

Another vision of the role of AI in art comes from the Bianjie (邊界) Research Group, an artist collective and research group working on the intersection of language, memory, and machine cognition. I spoke to Alexandre Montserrat about the work of the collective a few weeks ago, who showed me a few immersive videos of the installations. As you can probably tell from exploring the website, it is hard to describe the content, but the immersive video gave me a window into the future of art unlike anything I’ve seen before. The installation used video and AI-mediated feedback to construct an evolving symbolic infrastructure, simulating what it is like to be inside a human brain while it is associating different concepts and memories to form a thought. The work didn’t use AI as a content generator like the slew of Ghibli-style images that exploded on the internet after the release of GPT 4o. AI was a collaborator in creating an experience that would have been impossible without it—which was still, at its core, a deeply human expression of creativity. After reading a draft of this essay, Alexandre sent me the following post from Ian, his co-founder, which he said best captures his views on AI and art:

Both Mendez and Bianjie point to the same future: AI will be in art, it will be the art, but the artist will be right there, holding the (digital) paintbrush.

Art as the Destination

So do I think AI will affect art? Of course, it already has. But unlike with truck driving or software programming, the impact will be markedly different. AI will be like the camera: it will both make some forms of art obsolete, but at the same time create a whole new canvas of possibilities for the artist. It will lower barriers, expand possibilities, and create entirely new forms of expression that we just can’t imagine.

But the impact of AI on art will be larger than giving artists another tool: as the production frontier expands for other commodities—as AI makes it cheaper to produce commodities and services that do not have humanness in their value—we will also see demand shifting toward goods and experiences that we would otherwise consider art. Things where the “touchy feely” matters, where provenance, craft, story, and identity of the artist are central to the experience, e.g., live performance, artisanal furniture, “homegrown” produce, etc.

As more value becomes concentrated in products that contain an artistic dimension, we will get more artists. If there is anything the economist in me knows: incentives matter. Some day, when the machines are doing most of the stuff that we currently see as “work,” we may look around and see that others are engaged in activities that are both familiar and rare in today’s world—making things, performing, telling stories, crafting experiences for one another—i.e., what artists have been doing all along.

I’ll end with my collaborator Rohit Krishnan’s post on how he uses AI art, because I think it really captures the space that is opening up to us: