Case Study 20.2: The AI-Generated Painting

Background

The Meridian Prize is one of the most prestigious international awards in contemporary visual art — not for the largest sum (it is not), but for the distinction it confers. Previous winners include names that now appear in major museum collections worldwide. The prize draws entries from across the globe; its jury is composed of working artists, critics, curators, and art historians, selected to represent geographic and stylistic diversity.

This year, the jury deliberated for three days. The winner was announced: Passage III, a large-format work on the theme of grief, mortality, and the dissolution of self. The jury's statement described it as "formally rigorous, emotionally profound, and wholly original in its treatment of a theme that has attracted every artist who ever lived."

What the jury did not know — because it was not disclosed in the submission, and the competition rules did not require disclosure — was that Passage III had been generated by an AI image synthesis system. The human "artist" who submitted it, Nadia Kowalczyk, a thirty-four-year-old visual artist based in Warsaw, had spent six weeks developing and refining the text prompts that directed the AI system. She describes her process: "I wrote hundreds of prompts, discarded thousands of images, made choices about color, composition, tone, and concept at every stage. The AI was my medium. I am the artist."

The announcement triggered an immediate and intense controversy. Within forty-eight hours:

• Seventeen of the forty-three shortlisted artists issued a joint statement demanding the prize be rescinded
• The jury issued a statement saying they stood by their judgment: the work was beautiful and moving, and they had no regrets about their choice
• The competition organizers announced an emergency meeting to revise the rules for future years
• Three members of the jury quietly told reporters they felt deceived and would have voted differently if they had known
• Several major critics published pieces arguing on opposing sides
• The internet produced the usual volume of heat and almost no light

Nadia Kowalczyk gave one interview, then went silent.

The Philosophical Questions

Does the Origin of a Work Affect Its Aesthetic Value?

This is the central philosophical question, and different frameworks give different answers.

The Platonic answer would seem, at first, to favor Kowalczyk. For Plato, beauty is in the object (or rather, in the Form that the object participates in). The Beautiful Itself is what it is regardless of how it came into being. If Passage III genuinely participates in the Form of Beauty — if it genuinely embodies the formal qualities that constitute beauty — then its origin in an AI system is aesthetically irrelevant. The jury found it beautiful; it was beautiful; the end.

But Plato's position is more complicated. Recall his suspicion of mimetic art — art that works primarily through imitation and emotional manipulation rather than through rational engagement. An AI system, trained on millions of human-produced images, is a mimetic machine of unprecedented scale: it produces images that look like art without having any understanding of what it is producing. For Plato, the question might not be "is the image beautiful?" but "is this genuinely the appearance of beauty, or merely a sophisticated simulation of beauty?" He might argue that genuine beauty requires a rational agent who perceives and responds to the Forms — and that an AI system, however impressive its outputs, lacks the kind of engagement with the Beautiful Itself that genuine art requires.

The Kantian answer cuts more sharply against Kowalczyk. Kant's analysis of aesthetic judgment is focused on the experience of the audience, not the properties of the object — but his broader philosophy of art (especially in later sections of the Critique of Judgment) places significant weight on genius: the natural gift through which a great artist gives aesthetic ideas expression that cannot be reduced to any rule. For Kant, genius is not a technique that can be learned; it is a capacity for creative expression that exceeds any method.

An AI system, however sophisticated, operates by applying patterns derived from its training data. It does not have genius in Kant's sense; it simulates the products of genius. This suggests that even if Passage III produces the right aesthetic response in its audience, something is missing at the level of the work's creation — some dimension of artistic achievement that genuine art requires.

But here a complication arises. Kowalczyk's process was not simply "press generate and accept the first output." She spent six weeks making thousands of aesthetic choices — developing prompts, selecting and rejecting outputs, making decisions about color, tone, composition, and concept. Is this meaningfully different from a photographer who, through framing, lighting, timing, and selection, produces a beautiful image using a camera? A camera is also a tool that does not "understand" what it photographs; the aesthetic achievement belongs to the photographer who makes the choices. Why not to Kowalczyk?

The Deweyan answer offers what is perhaps the most useful framework for thinking about the controversy. Dewey's question is not "what is the origin of this object?" but "what is the quality of the experience it generates?" If Passage III genuinely produces a consummated aesthetic experience — if it draws audiences in, rewards sustained attention, generates genuine emotional and cognitive engagement — then it has achieved something aesthetically real. The jury's experience of the work, their sustained engagement and genuine response, is the relevant aesthetic fact.

But Dewey is also interested in the process of creation as an esthetic experience in itself. The artist who spends months struggling with a canvas, revising and rejecting and starting over, is having an esthetic experience as well as creating an esthetic object. The creation process is not just means to an aesthetic end; it is itself an esthetic engagement with the material world. Kowalczyk's process — writing prompts, making choices, iterating — is a form of esthetic engagement, but it is a different kind than the traditional artist's engagement with physical materials, with the resistance of paint, with the demands of scale and surface. Whether this different process disqualifies it as genuine art is, in Deweyan terms, a question about the character of the process, not about the origin of the output.

Can AI Produce Genuine Art?

The question turns on what we mean by "genuine." At least three distinct issues are in play:

Authorship: Who is the author of Passage III? Nadia Kowalczyk? The AI system? The human artists whose work was in the training data? There is no settled legal or philosophical answer. The copyright question (who owns the work legally?) is being litigated in multiple jurisdictions. The philosophical question (who is the artist in any meaningful sense?) is harder.

One useful distinction: the difference between procedural authorship (the entity that executes the procedures that produce the work) and intentional authorship (the entity whose intentions and choices shaped what the work is about and how it works). By this distinction, the AI system is the procedural author; Kowalczyk is the intentional author — the one who decided this work was about grief and mortality, who made choices about the aesthetic register and formal character, who decided when the process was finished. If authorship in the artistically relevant sense is intentional rather than procedural, then Kowalczyk is the author, and the AI is the medium.

Intentionality: Does genuine art require human intentionality — suffering, experience, care, the full weight of a human life behind the choices? Many of the protesting artists feel this way, and it is a philosophically serious position. Rasa theory might support it: the sahridaya's experience of rasa is, in part, a resonance between their emotional experience and the intentional emotional engagement of the artist who shaped the work. If there was no human emotional engagement in the creation of Passage III — only Kowalczyk's prompts and choices — is the rasa response in the audience a form of being moved by something that was never actually felt?

But this objection proves too much. We do not typically require that all art be the direct expression of the artist's personal suffering. A composer who has never experienced the grief she depicts in a requiem may nonetheless produce a work of genuine esthetic value. A novelist who has never experienced war may nonetheless write a novel about it that is truthful and moving. The requirement that the artist have personally felt what the work depicts is both too strong and too limiting.

Reception: The jury's experience of Passage III was genuine. Three of them may have said they felt "deceived" — but what they actually experienced in the three days of deliberation was not deception; they were genuinely responding to a work that was genuinely before them. The question of whether they would have voted differently had they known the origin is interesting: it suggests that their aesthetic response was conditioned by a background assumption (that this was made by a human artist with human intentionality) that turns out to have been false. Does this mean their response was not a genuine aesthetic response to the work itself, but to a story they were telling themselves about the work?

This is the hardest question. Kant says aesthetic experience involves the free play of imagination and understanding in response to the object itself. If what you are actually responding to is a narrative about the object (this was made by a suffering human being), and that narrative turns out to be false, is your aesthetic response genuinely tracking the object? Or are you responding to a fiction?

What the Debate Reveals About What We Value in Art

The controversy over Passage III is not primarily about aesthetics in the technical sense; it is about what art is for and what we actually value when we value art.

If what we value is formal and sensory qualities — the composition, the color, the emotional resonance of the visual field — then the AI-generated work can have these qualities in full measure, and the question of origin is irrelevant.

If what we value is human achievement — the discipline, the suffering, the years of training, the repeated failure and recovery that culminate in a great work — then AI-generated art fails on this dimension, because there is no human achievement in the creation of the visual output (whatever achievement there may be in Kowalczyk's prompting process).

If what we value is connection — the sense of being in contact with another consciousness, of encountering another human being's vision of what grief or joy or horror looks like — then AI-generated art faces a genuine philosophical challenge: there is no other consciousness there to connect with, at least not in the visual synthesis itself.

If what we value is the experience produced in the audience — the quality of the encounter between the work and the prepared, sensitive viewer — then the jury's deliberation is the relevant fact, and the origin is secondary.

Questions for Discussion

1. Was the jury right to award the prize to Passage III? Was Kowalczyk right to submit it without disclosure? These are separate questions — try to answer them separately.

2. Apply the three frameworks (Platonic, Kantian, Deweyan) to the question: can AI produce genuine art? Which framework provides the most useful analysis? Which is most limited?

3. The protesting artists argue that knowing Passage III was AI-generated should change how we evaluate it aesthetically. Is this a form of the "genetic fallacy" (judging a work by its origin rather than its qualities)? Or is the origin of art genuinely relevant to its aesthetic value?

4. Rasa theory holds that great art allows audiences to experience and savor emotions in a purified form, partly through the resonance between the artist's intentional shaping of the work and the audience's emotional receptivity. Does this framework provide grounds for excluding AI-generated work from the category of genuine art? Or can Kowalczyk's intentional choices supply the relevant intentionality?

5. What rule, if any, should the competition organizers adopt going forward? Consider: (a) require disclosure of AI use and treat AI-generated works as a separate category; (b) treat AI as a medium like any other and judge only the output; (c) disqualify AI-generated works entirely. Defend your choice using at least one framework from the chapter.

6. The "genuine art" debate seems to hinge on assumptions about what art is for. Based on your reading of this chapter: what do you think art is for? And how does your answer affect your view of whether AI can make genuine art?