AGENCY

We sometimes read a controversial paper that resonates with our recent experiences. In her article below Belinda Levez reflects on the inherent weaknesses of the contemporary art world and how they are being exposed by AI and other cultural stresses.

What AI reveals about value, labour, and structural weakness in the art world

Belinda Levez   May 03, 2026

Headlines scream “AI destroys art!” but almost none of them ask the uncomfortable question: is AI actually the problem—or is the panic just a convenient distraction from the structural weaknesses in the art market that have been ignored for years?

The current discourse around AI and creativity is loud, emotional, and mostly unserious. It treats a tool like a villain and frames a system-wide disruption as a crisis. But underneath the noise, something more accurate is happening: AI is not breaking art. It is exposing how fragile, inflated, and poorly understood parts of the creative economy already were.

The dominant narrative is simple: AI generates images, therefore artists are in danger. But that framing skips every meaningful layer in between. It assumes that art markets, creative labor, and cultural production were stable before AI arrived. They weren’t.

Long before AI-generated images entered the conversation, the creative world was already dealing with structural imbalance. There was an oversupply of artists trained to produce, and an undersupply of systems capable of supporting them meaningfully. More people were creating than there were mechanisms to evaluate, distribute, or sustain that creation.  AI didn’t create that imbalance. It just made it visible.  And visibility is uncomfortable.  So instead of examining the system, people blame the tool.

Oversupply was already the real story. There are far more artists than there is demand for traditional artistic labor. Design schools graduate thousands of students. Online platforms produce millions of self-taught creators. Social media rewards constant output. Courses and tutorials lower barriers to entry every year.  The result is not scarcity of art—it is saturation.

Even before AI, most artists were competing in a crowded, attention-starved environment where visibility mattered more than skill, and marketing mattered more than depth. Success was never purely about talent. It was about timing, narrative, access, and luck as much as it was about ability.  And the system’s filters have always been far narrower than people like to admit. Only a small fraction of artists ever pass through institutional gates with just six percent reaching gallery representation, and one percent entering museum-level visibility. The vast majority were never being “selected” in the first place.  AI does not fundamentally change that system. It accelerates it. The panic is not about a new problem. It’s about an old problem reaching its logical extreme.

History has already shown how this kind of panic unfolds. When early photographic processes like the Daguerreotype emerged, many artists—especially portrait painters—saw them as an existential threat. For centuries, accurate representation had been a core economic function of art. Photography automated that function almost overnight.  The reaction was familiar: dismissal, anxiety, and cultural resistance. Critics like Charles Baudelaire argued that photography was mechanical and soulless, a degradation of true artistic practice. And economically, the fear wasn’t irrational—some traditional roles did decline as cheaper, faster image-making took hold.

But photography didn’t destroy art. Once painting no longer needed to serve realism as its primary function, it evolved. Movements like Impressionism shifted toward perception and light. Later, Expressionism and Cubism pushed even further away from representation. Meanwhile, photography itself became an art form, championed by figures like Alfred Stieglitz.  The pattern is consistent: when a technology makes one layer of creation cheap, value moves somewhere else.  AI is not an exception to this pattern. It is an acceleration of it.

AI didn’t break the market—it revealed it. The art world likes to imagine itself as a meritocracy of taste and talent. In reality, it has always been a layered system shaped by access, narrative control, and gatekeeping.  Collectors don’t discover art in a vacuum. They rely on signals, institutions, and intermediaries. Galleries don’t simply find “the best work.” They curate within constraints: relationships, trends, branding, and market positioning. And artists are often forced to navigate a system where visibility is not proportional to skill, and success is not proportional to effort.

AI does not introduce corruption into this system. It exposes the fact that selection has always been partially arbitrary.  When generation becomes cheap, the illusion of scarcity disappears. And when scarcity disappears, the weaknesses in curation become obvious.  The myth of pure talent was always convenient. A lot of the current anxiety comes from the belief that traditional artistic success was purely based on talent and hard work. That belief was never fully accurate.

Skill matters. Dedication matters. But access, timing, narrative, and institutional alignment have always played major roles in determining who gets seen and who gets supported.  AI disrupts this not by making talent irrelevant, but by removing one of the last illusions of exclusivity: that technical production alone is what separates “real artists” from everyone else.  It turns out production was never the bottleneck people thought it was. Distribution always was.

The real pressure point is attention, not creation. One of the biggest misunderstandings in the AI debate is the assumption that the crisis is about making art. It isn’t. It’s about attention.  We are moving into a system where creation is effectively infinite, but attention remains finite. That shift changes everything.  When supply explodes and demand stays constant, the bottleneck becomes selection. And selection is not a technical problem—it is a cultural, psychological, and institutional one.  This is why AI feels destabilizing. Not because it replaces artists, but because it floods the environment where artists used to compete for visibility.  In a saturated system, the question is no longer “Can you make something good?” It becomes “Why should anyone care?

Collectors were not prepared either. It’s easy to focus on artists in this debate, but collectors and audiences are part of the same system—and they are equally unprepared for what AI reveals.  Many collectors rely heavily on external validation. They look to galleries, curators, social proof, and institutional signals to determine value. This isn’t necessarily a flaw; it’s how complex cultural markets function.  But it creates vulnerability.  When AI floods the market with high-quality visual output, the surface-level markers of “value” become less reliable. Aesthetic appeal alone is no longer enough to distinguish significance from noise.

This forces a harder question: how do you evaluate art when technical execution is no longer scarce?  The uncomfortable answer is that many collectors were never evaluating depth as carefully as they assumed. They were often responding to framing, context, and reputation as much as to the work itself  AI doesn’t destroy that system. It forces it to become more honest.

Galleries and gatekeepers lose their monopoly on filtering in the same way. For decades, galleries and institutions functioned as filters. They didn’t just showcase art—they defined what counted as art worth seeing.  That role gave them enormous influence, but also a kind of monopoly over visibility.  AI weakens that monopoly by changing the economics of production and distribution. When anyone can generate high volumes of competent work, exclusivity becomes harder to maintain based purely on output scarcity.

And that shift is already visible beyond traditional institutions. Many online platforms have begun restricting or outright banning AI-generated images—not as a solution, but as a reaction to the same saturation problem. When supply becomes overwhelming, systems default to blunt forms of control.  This doesn’t make institutions irrelevant. It forces their role to evolve from gatekeeping toward meaning-making.  Instead of asking “what should be shown,” the question becomes “what actually matters in a world where everything can be shown?” That shift is not optional. It is already happening.

The winners will not be the loudest critics. In moments of disruption, there is always a predictable pattern: those who feel threatened often become the loudest voices in opposition.  We are seeing that now. A wave of resistance frames AI as an existential threat to creativity itself. But refusing to engage with the shift does not slow it down. It only guarantees irrelevance.  The artists who will thrive are not those who resist tools on principle. They are those who adapt their thinking to new constraints.  That means moving away from purely execution-based identity and toward idea-based practice. It means understanding that output is no longer the scarce resource. Judgment is.

Evolution does not mean surrendering identity. There is a false binary in the current debate: either reject AI entirely or lose artistic integrity.  That framing is lazy.  Evolution in creative systems has never required abandoning identity. It requires redefining where value actually sits.  Photography did not kill painting. Digital tools did not eliminate illustration. New mediums consistently shift the boundaries of practice without erasing human expression.

AI is another shift of that kind—but faster, broader, and more destabilizing in its early phase.  The mistake is assuming adaptation equals compromise. In reality, adaptation is how relevance survives.

The real divide is not human versus machine. It is between those who understand systems and those who don’t.  Some artists will learn to use AI as an extension of their thinking, amplifying ideas, accelerating iteration, and exploring possibilities that were previously too expensive or slow to pursue.  Others will treat it as an external threat and remain stuck evaluating their worth based on a model of creativity that no longer matches reality.  The difference is not technical. It is conceptual.

The system was always the problem. If AI exposes anything clearly, it is that the creative ecosystem was never as stable or fair as it liked to believe.  Oversupply of creators, inconsistent valuation systems, opaque gatekeeping, and attention-driven economics were already defining the landscape long before AI arrived.  The technology didn’t break the system. It removed the ability to ignore its flaws.  And that is why the panic feels so intense. Not because something new has been destroyed, but because something old can no longer be rationalized away.

AI is not the enemy of art. It is a mirror reflecting how art systems already function under pressure.  It shows us that production was never the real bottleneck, that visibility was always unevenly distributed, and that value has always been shaped by more than just skill.  The discomfort people feel is not caused by AI alone. It comes from realizing that many assumptions about creativity, merit, and success were incomplete.

The winners in this shift will not be those who resist change. They will be those who look at it clearly and adjust accordingly: artists who evolve, collectors who learn to evaluate beyond surface signals, and institutions that move from gatekeeping to meaning-making.  Everyone else will be left arguing with the system while it moves on without them.