Urinals and AI in Art

Last updated: June 20, 2025

Background

Machines are embedded in art so deeply that we don’t even really notice them. Most writers don’t think it’s a perversion of their work to type it on a computer. Even the most conservative likely doesn’t deactivate the spell check built into Google Docs or Microsoft Word. But that same writer probably rebukes ChatGPT and regards AI-produced writing as second class, in hypotheticals and when they can spot it. They complain with their friends and at conferences about how it makes bad content and defiles the craft and yet is slowly but surely displacing creators in their industry. Fundamentally, a lot of artists don’t regard AI-generated art as art. Some instances of engineering in art are acceptable, though; they might defend the honor of an artist who uses a tool like Procreate to create animations and drawings. They appreciate the daring of a readymade sculpture.

As the conversation around AI art becomes louder and more frantic, especially after the launches of Veo 3 by Google and 4o Image Generation by OpenAI, I thought it would be interesting to look at Marcel Duchamp’s Fountain (1917), a piece that altered the course of modern art history and forced an expansion of the criteria that define art. Right now, the definition of art is being challenged in some fundamental way - it seems to mean something new and different for algorithms to make art products that we consume. Our social media feeds, ads, and even galleries are being populated with them rapidly. But the academic definition of art has changed many times. Engineered products aren’t new to art by any stretch of the imagination. So what’s actually changing now? Is this a bigger and more fundamental shift than those that have come before? How does the intrusion of AI into art making fit into the historical progression of modern art? How does it compare to the shockwaves felt post-Fountain?

Fountain by Marcel Duchamp (1917)

Fountain is a porcelain urinal signed “R. Mutt”, submitted for the inaugural exhibition of the Society of Independent Artists with Duchamp’s assertion that it is an everyday object “raised to the dignity of a work of art by the artist’s act of choice”. Fountain wasn’t displayed in the show area of the exhibit and the original has been lost. It ended up being displayed in Alfred Stieglitz’s studio and published in The Blind Man.

This work now lives at the heart of the 20th century discussion around what constitutes art, “which more than any other experiment has challenged the boundaries and even the foundations of art as a concept” (Goldsmith). Although Fountain was initially met with chuckles and repulsion, it ultimately was accepted into museums and displayed on a pedestal, as Duchamp had intended. It locked itself into the course of modern art history and prompted a wave of celebration of everyday objects, sometimes as commentary on consumerism and advertising. For example, Andy Warhol’s iconic Campbell Soup Cans (1962) and Claes Oldenburg’s Shuttlecocks (1994).

Relationship to AI

The act of creating an art product with generative AI is reminiscent of Duchamp’s process. The user inputs a prompt - their will - to a chatbot, which returns an image. Is the image art, considering that Fountain is? There doesn’t seem to be a huge difference between writing a prompt and receiving a visual representation of it as compared to having a thought and applying it into a pre-engineered urinal and saying that it was your unique human brain that justifies its presentation in a museum.

I’m sure there are artists and art critics who would disagree with that comparison. The consensus seems to be that Duchamp’s originality was in realizing that there was something to be said about what we accept as art. A person who uses an image generator isn’t doing something for the first time. They’re using it as a tool to circumvent the great struggle associated with developing and applying one’s technical art skills to a paper or block of clay. They’re using it to avoid figuring out every detail of the output, another key element of art of the past. There are a lot of romantic notions about how a painter knows everything that goes onto their canvas and that everything has a purpose. Chekhov’s gun lalala. But AI might put something in your output that you didn’t explicitly ask for or intend. The AI-generated gun won’t necessarily be fired, which defies our expectations as viewers.

Richard Kuhns, in Art and Machine (1967), writes that Duchamp’s “discovery” of the readymade introduced two big ideas into art - the engineered itself as an art form and chance as artistic choice. AI in art certainly allows for an appreciation of both the engineered and chance. When we see a startlingly real-looking live comedy show, we don’t praise the prompter of the video, we praise Gemini’s engineering triumph. Complete precision is impossible - the user doesn’t prompt every pixel. If you ask ChatGPT to paint you a blue sky without specifying the shade, you leave it up to chance. You still imagine a blue sky, but you don’t determine exactly what it ends up looking like.

Is AI just a tool?

As of now, sort of. NVIDIA’s AI Art Gallery site says that artists are “generative AI as a tool, a collaborator, or a muse to yield creative output that could not have been dreamed of by either entity alone”. In the way that there are painters and sculptors, a new type of artist is emerging - one whose medium is these models. But because AI labs are aiming to create agents not reliant on human intervention, any description of AI as a mere “tool” seems, if not outright deceptive, temporary.

A big part of what people found disturbing about Fountain was the complete lack of physical intervention by the artist. Duchamp didn’t do much except have the idea, sign the urinal, and show it to people. AI also enables the creation of art with barely any effort on the user’s part. It can do an artist’s whole job nearly end to end. The necessity of writing a prompt and making edits are really the only steps that make AI permissibly describable as a “tool”. But as these agents get better, they’ll be able to eliminate, or at least corrupt, the act of thinking as a step in creation. Don’t know exactly what you want to make? Prompt an LLM for a prompt! AI, if it can’t already, will be able to rob the artist not just of the technical effort of creation but the thinking that they treasure so deeply. It was human thought that made Fountain into a “sculpture” and induced one of the most important inflection points in modern art history - when both the effort of creation and the artistic intentionality aren't human, what’s left?

Anti-AI purism and reimaginations

Tim Schneider’s recent Substack post critiquing Sam Altman’s blog post “The Gentle Singularity” argues that ideas and production volume run counter to what makes art great. He writes that we don’t need more art - that we were already drowning in it before generative AI made its production so effortless and thoughtless. He quotes Neal Stephenson on a 2024 podcast.

“The real purpose of art and the reason we like art is because it exposes us to a very dense package of micro decisions that have been made by the artist. As such, we’re engaged in a communion with that artist. What makes it interesting is that connection. It may be to a living writer, or it may be to a sculptor who died 2,500 years ago, but in either case, we’re making a human-to-human connection. If we know that we’re reading something or experiencing a work of art that was just generated by an algorithm, then that element of human connection isn’t there anymore.”

Tiny choices are what drive us to seek art. AI lets the artist make big choices rather than small ones, hollowing out the meat of the artwork that Schneider argues that the viewer is looking for. There are surely thousands of depictions of the crucifixion: the same scene painted or sculpted again and again by artists across centuries and cultures, differentiated only by minute artistics decisions of how to proportion the body on the cross, which characters to include, and the level of agony to portray on Jesus’ face. These choices make each painting vastly different from the next, even though the story is the same.

AI-free art will never die; some people will always have the urge to create on their own. Artists didn't stop painting after the invention of the camera and didn't stop sculpting after the emergence of the readymade. Viewers will always seek art that serves as a bridge to the soul and unique experiences of the artist. But artists, increasingly, will learn to incorporate AI into their work to extend their knowledge base and speed up production - Hard Fork recently brought Pete Wells on to talk about how chefs are using chatbots to find new recipe ideas, similarly to how they might ask their sous-chefs to find inspiration. The kind of art that artists sell to clients and submit for grading in schools - graphic design, video creation, and writing, for example - are all just a prompt away from the people who used to pay humans for those services. And viewers undeniably enjoy perfectly optimized content, especially when they can't quite tell that what they're seeing is AI - TikToks users are apparently having a field day with AI-generated ASMR content.

I didn’t use AI for this essay because I wanted to read for it and think about it. It took way longer than it would’ve if I had asked an LLM for ideas or to restructure my notes. I like the struggle associated with writing and knowing that this all came from my head. But this is also just for fun and I'm not trying to sell this to you or even perfectly answer my own questions. I didn't want to dirty up my thinking by putting it through a chatbot, which I might've felt differently about if this was a different piece of writing for a different purpose. The resistance among artists to letting AI “corrupt” what they do is strong but complicated - a combination of anxiety about their livelihoods and a moral opposition to shortcutting artistic traditions, qualified by the reality that AI can, in at least some ways, make their work faster. I think that resistance isn’t new, but a descendant of the existential tussle with machinery in art that came into the center of discussion after the creation of Fountain in 1917.