@SHL0MS is an anonymous artist whose practice spans performance, digital art, and conceptual experiments that consistently catch the internet off guard.
The experiment began when SHL0MS posted a cropped version of Claude Monet's Water Lilies on X, framing it as AI-generated and asking followers to explain what made it inferior to a real Monet. Over 600 responses poured in, with confident takes citing missing impasto, unnatural reflections, and a lack of soul. The experiment went viral, and SHL0MS minted the painting as an NFT titled "Inferior Image," which sold for just over $40,000 across 28 bids.
In this interview, SHL0MS discusses what the flood of responses revealed about motivated perception, why the AI label acts as social permission to bypass critical thought, and how the NFT is less about the image and more about owning the story.
OpenSea: What made you choose this specific Monet, the 1914-1917 Water Lilies series, as the subject of the experiment? Was there something intentional about picking a work that's so iconic?
SHL0MS: My main criterion was that it not be instantly recognizable. I wanted something where the viewer would have a vague idea of "Monet" in their head but would not be able to recognize this Water Lilies on sight. There are 250 of them, which means the average person’s mental image of what makes it good or what it looks like is far more abstract than something like Starry Night. I also did not want a wide-format Grandes Décorations piece because I did not want to crop a minor detail of a larger work. It also needed to be high-resolution and reverse searchable, so the more skeptical members of the audience could identify it easily. The one I ended up with was a nice high-resolution scan from Wikimedia. I cropped a small amount on the left hand side to remove the signature.
The iconic factor mattered in the inverse direction. I wanted something Monet-adjacent enough that people would feel competent critiquing it. If I had used an obscure painter nobody had heard of, people would have been more careful; the Monet label gave them license to be confident.
OpenSea: When you posted the piece on X, did you expect over 600 people to respond, or were you surprised by how confident and detailed the critiques were?
SHL0MS: The volume did not surprise me so much as the sheer confidence. The practice is largely improvisational; I do not know in advance whether something will go viral, and most things do not. Volume is unpredictable on its own; virality is a bit of a collaboration between human and algorithm that is equally skill, art, and random luck.
What surprised me was how detailed the critiques were and how willing people were to deliver them with an air of assumed authority. There were actual painters and art critics among the people who satisfied the criteria, but also many who, by any reasonable standard, had no business issuing an aesthetic verdict on a Monet. The AI label provided social permission to let the mask slip. Once they thought it was AI, the epistemic standard for confidence or critical thought collapsed.
OpenSea: The experiment surfaced something uncomfortable: that people will construct elaborate, technical-sounding arguments to support a flawed premise. What does that tell you about the way we talk about art right now?
SHL0MS: We have built a vocabulary for dismissing AI work that we deploy automatically the moment we think we are looking at AI. The vocabulary has a technical-sounding quality. Incoherent brushwork, dead composition, no soul.
This is a story about motivated perception more than about AI. We have always done it with categories we have opinions on; AI is the category where the failure mode happens to show up cleanly right now. People were quick to point out that this is not a new phenomenon, and I agree. The point is to reveal irrefutably that the phenomenon is happening now.
OpenSea: Do you think the result would have been different if you had framed it as a lesser-known human artist's work rather than AI? What specifically is it about the AI label that triggers that response?
SHL0MS: Art gets enshrined in history through a wide variety of factors. The Mona Lisa is famous because it was stolen. I think that today, with so much anxiety and confusion about the role of authorship, people are driven by defensiveness to make objective assessments on aesthetics for work which is more historically constructed than they’d like to admit. We are right between the old paradigm where the main critique of AI generations was that they were incomprehensible and ugly and could never produce a human with the right amount of fingers, and a new paradigm where it was a much more intrinsic spiritual lack of soul but the generated image can indeed be perfect on a technical level.
Framed as a lesser-known human artist, I do think people would have been more cautious. Famous artworks and artists have a religiosity to them, a fundamental value that cannot be challenged. They are the framework through which the public evaluates all other art. They serve a load-bearing, epistemic purpose in people's understanding of beauty and ugliness and genius and creativity and what it means to be human.
The AI label tells you it is safe to dismiss the work without looking at it: the work was made by something we have already decided is suspect. The religious fervor around the original work is what triggers the ferocity and certainty of the response.
OpenSea: There's an irony in the fact that the "flaws" people identified, like unnatural reflections, are real qualities that exist in the painting. How do you think about the relationship between bias and perception in how we look at art?
SHL0MS: The work changed my own relationship with the painting. Someone pointed out a garish purple outline around the lilies as evidence the image was AI. I cannot unsee it now. They were right, that it is there. They were just wrong about what put it there. The perception was accurate; the framing made it a flaw.
We see what is on the canvas. We just attribute it differently depending on what we already believe about the source. An AI's garish purple outline is sloppy; Monet's is a bold use of color. The pigment is the same.
OpenSea: Minting the painting as an NFT feels like a deliberate extension of the experiment. And naming it "Inferior Image" adds another layer to that. What were you trying to say with those choices, and who do you think will want to own it?
SHL0MS: The NFT was not the artwork. I have minted maybe a few dozen NFTs across nine years of practice. Some are standalone artworks, and some are pointers on the blockchain that legitimize digital performance work that occurs offchain for an audience that can only relate to art through ownership.
The title points at the audience's reflex. Once they thought it was AI, the rubric of inferiority was the rubric they brought to it. The Monet was not inferior. The way of seeing was. The title lets them complete the loop on themselves; whoever buys it is buying a pointer to that loop.
OpenSea: You work at the intersection of technology, provocation, and art. Do you think AI is changing what we value in art, or just exposing biases that were always there?
SHL0MS: Both, but the exposing part is more interesting and more durable. The biases AI is surfacing are not new. We have always done motivated perception, in-group/out-group reading, deference to category over object. AI is a particularly clean test because the category is so emotionally charged right now that it overrides almost any other perceptual input.
The surface of the discourse is shifting. The old critique of AI was mechanistic: we can tell, because the system cannot do the thing. Will Smith eating spaghetti, six fingers instead of five, wine glasses that defy physics. Now people are granting AI capabilities it does not have and falling back on "no soul." The critique has moved from mechanistic to ineffable, which means it is no longer testable. That shift is happening fast. Both framings (AI changing what we value, AI exposing what was always there) capture part of what is going on.
OpenSea: What do you want people to actually take away from this? Is it a critique of AI skepticism, of art world gatekeeping, or something else entirely?
SHL0MS: I am not interested in handing the audience a neat takeaway. The work is a demonstration rather than a critique, neither of AI skepticism nor of art-world gatekeeping. The underlying claim, that we reflexively classify and dismiss based on category rather than object, is uninteresting as a stated proposition; a thousand people proving it on themselves is what makes it interesting.
OpenSea: How do you think about what makes art good versus what makes it important?
SHL0MS: I think that art needs to change something in you. It can change your mind, or the level of abstraction you operate on, or the way you feel, or the way you see a lilypad.
I also do not think importance and goodness are the same thing. Some of the art I think is good is small and almost nobody sees it. Some of the art I think is important is not great as object. Inferior Image is closer to the second category. The Monet is closer to the first.
OpenSea: The NFT you minted to commemorate the experiment sold for over $40,000. How did that make you feel? What do you think that says about the moment and the way people value digital art and digital ownership?
SHL0MS: The sale was controversial in NFT circles, mostly because it was a high 1/1 sale during a dry period but also because it was an easy piece for artists to compare with their own and feel like it took less effort. I got accused of money laundering, which is the standard NFT-world response to any sale somebody does not understand. I am of course not laundering money, and if I were, I would not do it through a viral artwork on a public immutable ledger. The buyer, @Jediwolf, is one of the largest collectors of historical AI artworks and we had never interacted before this piece. He was tweeting during the performance and contextualizing it and actually got more engagement on his posts than mine did, all before he knew there was going to be an NFT involved.
He truly just loved the performance and wanted to be involved, and ultimately own it. I’ve never had a collector jump in to shape a work in collaboration like that, it was quite an experience.
In an interesting way, some of the people who mint NFTs within a coherent genre for a living have a hard time viewing anything outside of that legible frame (a piece of media as the artwork itself) as legitimate, which is a strange reflection of the gatekept traditional art world that they seek to reject. The NFT is not the artwork; the artwork was the saga: the tweet, the replies, the curation of the critiques, the discourse layer on top of all of that. The NFT is an ownable encapsulation of the saga, a capstone. If the capstone clears at $40k during a market that has been pricing most things near zero, that says more about the buyer's read on the saga than about the value of the file.
People have stopped pretending the JPEG is the essential thing. The JPEG is a pointer to the story, and the story is what they are buying. That is closer to how art has worked the whole time. The mistake of early NFT culture was acting like the file was the artwork. Most of it was not, and the pieces were always pointing at something else: provenance, community, reference, joke. This case is no different. The art is the story.
Disclaimer: This content is for informational purposes only and should not be construed as financial or trading advice. References to specific projects, products, services, or tokens do not constitute an endorsement, sponsorship, or recommendation by OpenSea. OpenSea does not guarantee the accuracy or completeness of the information presented, and readers should independently verify any claims made herein before acting on them. Readers are solely responsible for conducting their own due diligence before making any decisions.
