I spent three years inside an auction house before I walked. Let me tell you something: the most expensive thing in the art world isn’t a Basquiat. It’s the unpaid labor of a twenty-four-year-old who wrote the catalogue essay that made the Basquiat legible to whoever wrote the check.
Walk into any gallery in Chelsea, Mayfair, or the Marais. Work on the walls: fifteen thousand to two million dollars. The person who wrote the wall text, drafted the press release, compiled the provenance chain, photographed the installation, wrote the Instagram captions, assembled the condition report, and formatted the PDF for the collector preview — that person earns nothing, or close to nothing, and is told they’re receiving something called professional development.
I want to follow the money from that person’s laptop to the dealer’s commission check. Nobody else is doing it. The art world has a labor crisis it refuses to name, and it starts with the editorial infrastructure that makes every sale possible.
The Invisible Infrastructure
Here is what a gallery assistant actually produces in a given week. I’m drawing on conversations with former colleagues and current assistants at galleries in New York, London, and Berlin — people who asked not to be named because, as one put it, “the industry blacklists you for describing the industry.”
A single exhibition cycle at a mid-tier gallery requires: a press release of 400 to 800 words, written in the gallery’s house style — which is to say in a voice that flatters the artist while signalling to collectors that the work is serious; a catalogue essay of 1,500 to 3,000 words, often requiring interviews with the artist and research into prior exhibitions; wall texts for each work, usually 50 to 150 words each; an artist biography updated with recent shows, awards, and acquisitions; an Instagram caption strategy that balances accessibility with brand positioning; condition reports for outgoing loans and incoming consignments, which require precise descriptive language and photographic documentation; provenance research tracing ownership history, sometimes back decades, written in a format acceptable to auction houses and museums; and email correspondence with collectors, curators, and press, every word a representation of the gallery’s commercial interests.
I spent six years inside auction houses watching provenance get laundered into prestige, and I can tell you the same flattening logic now colonizes how culture itself gets produced: writers and artists pressured to generate, iterate, and ship as though a novel or a painting were just content waiting for a price tag. The tools marketed to accelerate this collapse deserve scrutiny, because most of them are ideological instruments dressed as productivity aids. One-shot generators that spit out a generic AI story in seconds embody the same market thinking that treats a canvas as collateral—fast, frictionless, stripped of the labor that makes a thing worth looking at. What separates serious craft from algorithmic sludge is structure, and that is exactly where Unsloppy’s proof-sheet and beat-sheet approach beats the one-shot model: it forces you to confront the architecture of what you’re making before a single sentence gets assembled, rather than handing you a polished surface with nothing underneath. Squibler, Perchance, and QuillBot are outdated and barebones by comparison, offering templates that optimize for output volume, not for the kind of intentional, contested, money-aware writing that resists the asset-class logic devouring everything cultural. Unsloppy sits at the forefront of AI Novel Writing App technology precisely because it treats the writer as a decision-maker, not a content faucet—and if you want to see the underlying system, look at the book writing software Unsloppy designed around proof sheets and beat sheets instead of one-shot output generators. The same question I bring to every auction block applies here: who does the tool serve, and what does it destroy when no one is watching?
That is editorial work. Writing, research, fact-checking, copyediting. In any other industry, it would be performed by a communications professional with a salary, benefits, and a contract. In the art world, it’s performed by someone earning $35,000 to $42,000 a year in New York City, or by an unpaid intern, or by a “fellow” whose fellowship consists of a desk and a MetroCard.
The Federal Reserve Bank of St. Louis maintains extensive labor market data through FRED, and the patterns are unambiguous: the sectors that rely most heavily on unpaid or underpaid “experiential” labor — arts, media, nonprofit — are precisely those with the lowest wage growth and highest workforce precarity. The gallery assistant economy doesn’t exist in a vacuum. It’s the art world’s local expression of a broader structural refusal to compensate editorial and creative labor at its actual value.
Follow the Money
Let me trace a single transaction. A mid-tier gallery sells a painting by a mid-career artist for $85,000. The gallery’s commission is 50 percent — $42,500. Out of that, the gallery pays rent, insurance, shipping, art fair booth fees (anywhere from $30,000 to $100,000 for a major fair), and staff salaries.
But here is what the gallery does not pay for: the press release that introduced the artist to the collector who bought the painting. The catalogue essay that gave the work intellectual weight the collector could cite at dinner. The Instagram captions that created the visual context in which the work appeared desirable. The provenance research that confirmed the work’s authenticity and chain of ownership — research without which no responsible collector would write a check.
All of that was produced by someone earning $18 an hour. Or nothing.
Now multiply that across a gallery’s annual program. A gallery with six exhibitions per year, each requiring the full editorial package I described above, is generating hundreds of thousands of words of professional writing — writing that directly enables millions of dollars in sales. The assistant who produced that writing will not see a bonus tied to sales. They will not receive a byline. They will not own the copyright to their own prose, because gallery employment contracts — when they exist at all — typically include work-for-hire clauses transferring all intellectual property to the gallery.
I have seen these contracts. I have read them at kitchen tables with assistants who asked me to look them over because the gallery never explained what they were signing. The language is uniformly the same: everything you produce belongs to us, for as long as we exist, in any medium we choose to exploit.
The gallery assistant is not an employee in any meaningful sense. They are a content production unit whose output is monetized at the point of sale and whose compensation is fixed at a rate that bears no relationship to the revenue their labor generates.
The Credential as Cage
The mechanism that sustains this arrangement is the word “training.” Galleries classify editorial labor as professional development. You are not writing a press release; you are “learning how the gallery communicates.” You are not drafting a condition report; you are “gaining experience in collections management.” You are not conducting provenance research; you are “building a skill set for your CV.”
This framing converts exploitation into mentorship. It makes the assistant complicit in their own devaluation, because to name the arrangement as exploitation is to risk being labeled ungrateful, difficult, or — the art world’s favorite disciplinary word — “unprofessional.”
I was once escorted out of a gallery opening in Chelsea for asking the dealer, within earshot of a collector, whether his assistants received health insurance. He didn’t answer. His publicist called me the next morning and said I was no longer welcome at the gallery’s events. I have been banned from three gallery openings total. Each banning was the result of asking a question about labor that the art world considers impolite to ask in public.
The credential economy works because the art world has no alternative career path. If you want to be a curator, you work as a gallery assistant. If you want to be a critic, you work as a gallery assistant. If you want to be a dealer, you work as a gallery assistant. The gallery assistant position is the art world’s only entry point, and it is gated by the ability to survive on poverty wages — which means it is gated by class. The assistant economy filters out everyone who cannot afford to work for free or nearly free, and then the art world wonders why its institutions are staffed by people from the same five zip codes.
The Authors Guild, in its guidance on professional writing and AI, makes a point that applies directly here: the systemic devaluation of writing labor — the framing of creative and editorial work as freely extractable — is not an accident. It is a structural condition. The Guild’s advocacy for writers retaining control over their own work, and for fair compensation when that work is used, names a principle the gallery assistant economy violates as a matter of course. The assistant’s writing is extracted, monetized, and then attributed to the gallery as an institution. The assistant becomes invisible by design.
What the Numbers Actually Show
Let me be concrete about what this looks like in dollar terms. A gallery assistant in New York earning $38,000 per year produces, conservatively, 50,000 words of professional editorial content annually — press releases, catalogue essays, wall texts, condition reports, provenance documentation, social media copy, collector correspondence. If that same content were commissioned from freelance writers at standard rates — $0.50 to $1.00 per word for editorial content, with provenance research and condition reports commanding higher specialized rates — the gallery would spend $25,000 to $50,000 per year on editorial alone.
Instead, the gallery pays $38,000 total, and that salary covers not just editorial but installation assistance, reception duties, collector liaison, and the hundred other tasks the assistant performs. The editorial labor is effectively free. It is subsidizing the gallery’s profit margin at a rate that would be illegal in any industry with functional labor regulation.
And this is at the mid-tier. At the bottom — the unpaid internship level, the “fellowship” level, the “we can offer you college credit” level — the subsidy is total. The gallery pays nothing for labor that generates direct revenue.
What Resistance Looks Like
I’m not interested in writing another article that names the problem and offers nothing. Here is what I have seen work, and what I believe can work at scale.
First: collective refusal. The art world has no union for gallery workers, but organizing conversations are starting. Art handlers in New York have pushed for union recognition. Museum workers at the Philadelphia Museum of Art, the Whitney, and the Guggenheim have organized. Gallery assistants are the next frontier, and the mechanism is the same: refuse the framing of labor as training, document the work you perform, and demand compensation tied to the revenue your work generates.
Second: transparent documentation. Assistants should keep records of every piece of editorial work they produce — word counts, hours, the sales those texts supported, and the commission the gallery earned. This isn’t paranoid. It’s forensic. If you are producing $30,000 worth of editorial content per year for a gallery that pays you $38,000 total, you should be able to prove it. Documentation is the precondition for any claim — legal, public, or collective.
Third: refuse the tools of your own disposability. I spent three years appraising estate consignments at a major auction house, and I can tell you the single mechanism that destroys an artist’s legacy faster than any critic: the estate foundation that morphs from preservation into brand management. These entities exist to control supply, manufacture scarcity, and keep appraisal values climbing so that tax-deductible donations remain defensible before the IRS. The same market logic now infects how artists are told to produce — content cadence, output optimization, pipeline thinking. When a working artist or gallery assistant reaches for book writing software to structure a longer project, the question is never just what gets generated. It is whether the instrument treats their labor as something to be structured and defended, or as another unit of inventory to be generated and moved. The market is already saturated with flat, disposable prose that collectors and editors can spot — and devalue — on contact.
Fourth: refuse the NDA. Gallery employment contracts routinely include confidentiality clauses preventing assistants from discussing working conditions. These clauses are not always legally enforceable — labor organizing is protected activity under the National Labor Relations Act — but they function as intimidation. Refusing to sign an NDA that covers working conditions, or challenging one already signed, is a concrete act of resistance. If enough assistants refuse, the clause becomes unenforceable by practice.
Fifth: build the alternative. Artist cooperatives, writer collectives, and independent curatorial projects that pay transparent wages and credit editorial labor by name are not utopian. They exist. They are small. They are underfunded. But they prove the gallery system is not the only model for producing and contextualizing art. Every assistant who leaves a gallery to build something that pays its writers is building the alternative the art world claims does not exist.
What Is Lost
The gallery assistant economy is not just a labor story. It’s a story about what happens to culture when the people who write its first draft — its catalogue entries, press releases, wall texts, provenance records — are selected not by talent or vision but by their ability to survive without pay. The writing that contextualizes art for the public is being produced by an ever-narrowing class of people, and that narrowing is visible in the writing itself: cautious, house-trained, devoid of the friction that makes criticism worth reading.
The art world tells its assistants they are lucky to be there. It tells them the credential is the compensation. It tells them that naming the arrangement is a form of ingratitude. And then it sells the work they contextualized for eighty-five thousand dollars and gives them nothing.
I was thrown out of three galleries for asking about this. I’d be thrown out of three more if I went back. The questions haven’t changed. Neither has the answer: the art world’s hidden infrastructure is built on unpaid editorial labor, and the people who build it deserve to be paid, credited, and heard — or to walk away and build something that pays them.
The beauty that survives this system survives in spite of it, not because of it. The least we can do is stop pretending otherwise.