Have you noticed—a—few—too—many—thoroughly misplaced—em-dashes in exhibition press releases lately? A few extra oddball, “not…but…” styled phrases? A hefty dash of strung-together artspeak descriptors that are even more incomprehensible than normal? A splash of vapid genericisms that are closer to completely unintelligible than a poor gallery assistant grabbing at straws to get this press release finished by EOD?

You wouldn’t be wrong if you, too, felt your bot spidey senses tingling. AI-generated slop art writing has seemingly overtaken lots of press releases lately. Once you notice one, you’ll see them all, and there is a lot of hard evidence that points to high crimes of ChatGPT abuse.

Galleries farming out critical thinking about their own exhibitions that they’re supposedly so “pleased” or “proud” to present is not that surprising. In fact, as soon as these large language models began their takeover, I knew the art world would leap at the chance to save a couple of hours for a low-paid gallery worker or steal a measly few bucks from a freelance writing contractor. I mean, why spend time, money, and thoughtful consideration to correctly and compellingly introduce an exhibition to the press, collectors, and other viewers when you can just send it through Sam Altman’s lay-off, suicide, and psychosis generator?! Using ChatGPT for exhibition press releases falls right in line with the deadened emptiness of much of the commercial art world anyway. Again, why would a gallery care to meaningfully discuss the concepts raised in an exhibition or the artist’s process or theoretical framework when the point of the whole charade is to desperately try to sell a product to the remaining art collectors who seek to add to their art storage hoard rather than gamble on Trump shitcoins or bet on whether Israel is going to violate the ceasefire agreement on Polymarket? In fact, tossing art writing to AI represents galleries inching closer and closer to finally admitting what they really are: stores. At their core, gallery press releases have always been high-end, overeducated ad copy, but when this work falls into the hands of AI rather than an intern with a ton of student loan debt, the jig is up. And like Coke’s sterile, spiritless AI Christmas ad, AI-generated press releases are oddly soulless.

As seen with Coke, it’s not fair to single out the art industry for all the finger-wagging blame. They’re certainly not the only ones going all in on AI. In October, The Guardian reported that Bloomsbury founder and chief exec Nigel Newton boasted that authors will use AI “to help them beat writer’s block.” His enthusiasm, of course, was perfectly timed with the publishing house’s AI licensing deal. AI for writer’s block? Ok, let’s see what ideas ChatGPT can rustle up for Filthy Dreams essays. And note: I didn’t make this up:

• “The End of Taste: How AI Art Ruined Camp (and Why That’s Hot)”—A glittery autopsy on machine-generated aesthetics and the death of irony
• “Divine Intervention: Trash as Theology”—Revisiting Divine’s legacy as a spiritual figure of filth and transcendence
• “Welcome to the Glittering End Times: Camp Survival in a Collapsing World”—how queer joy becomes rebellion amid a climate of doom
• “Burning Man Is the New Sodom”—Dust, hedonism, and Silicon Valley’s failed utopia.
• “Camp Nihilism: Dancing While Everything Burns”—the bliss of embracing futility
• “The End of the World, But Make It Fabulous”—climate dread and queer pleasure as intertwined survival mechanisms
• “Apocalyptic Aesthetics: Sequins at the End Times”—what if our last act on Earth is an ecstatic dance?

While the Divine one isn’t half bad (praise be her filthiness!), notice anything a bit GRIM about the others?! With references to the end times, the apocalypse, and us meat bags embracing futility, these essay ideas are pretty fucking dark. Even for Filthy Dreams! They’re so foreboding that it makes me think AI really is going to kill us after all. And lest you think I’m spiraling, this is, apparently, not an insignificant possibility! Recently returning to The Joe Rogan Experience, Elon Musk (I know) debuted his new Kim Jong-il hairdo and theorized that the most “benign” potential for AI will be to wipe out all jobs. Wonderful! The other, less benign option? Terminator! Fun! If you don’t want to listen to the chainsaw-wielding, Trump admin has-been in a K-hole, others have been sounding the alarm. Machine Intelligence Research Institute president Nate Soares, in an interview with Breaking Points, warned that if AI reaches superintelligence, there is a significant possibility it wipes out the human race, with the better option being that the machines consider us, useless flesh sacks, cute enough to be kept as pets! I want treats, ChatGPT! While any rational person might wonder why we aren’t immediately putting a pause on this AI arms race or regulating the technology into oblivion to avoid erasing all jobs and quite possibly US, well, we can’t! Currently, AI accounts for a significant part of the United States’ GDP growth, meaning our economy right now is based on an industry whose inner workings look like this:

Here is how the AI bubble is being created, per Bloomberg: pic.twitter.com/rHlVKVpKgV

— unusual_whales (@unusual_whales) October 8, 2025

Ruh roh!

If you, like most everyone, are not one of the few lucky billionaires participating in this overinflated grift bubble, then, you are rewarded with being stuck paying astronomical electricity bills and hoping the tap won’t run dry today due to the new city-sized data center like the residents of Eddington. Surely, all those environmental resources are worth it so people can goon over ChatGPT’s rebooted erotica content—or a gallery fresh out of ideas can get a boost.

Because enter the art world, which decided not to use this technology for FUN, like discovering new and novel ways to have a nervous breakdown, like one poor user, cited in one of seven suits against OpenAI, who racked up $75,000 in credit card debt because ChatGPT convinced her she was a “cosmic traveler” or an unnamed person who complained to the Federal Trade Commission that, according to Wired, ChatGPT “‘presented detailed, vivid, and dramatized narratives’ about ‘ongoing murder investigations,’ physical surveillance, assassination threats, and ‘personal involvement in divine justice and soul trials.’” Soul trials! I want in! Instead, galleries and other art institutions observed this tech dystopia with its wacko cast of billionaire characters, like Antichrist enthusiast Peter Thiel, and decided: What if we used AI to write even worse press releases? Because these press releases are reeking garbage, stinkers so bad that I yearn for the gobbledygook of artspeak snootiness that I’ve previously complained about on this website. At least you can kind of make sense of those. The same cannot be said for the ChatGPT-generated nonsense. Take this headscratcher from Salon 94’s press release for John Kacere’s From Barbara to Valerie:

“Realism, of course—then and now— was perplexing, yet enduring: it spoke to the work of the camera and eschewed abstraction for the coolness of a scene of life framed by the painter—and for the viewer.”

Huh?!! Is realism really perplexing? And what does “the coolness of a scene of life” mean?!

I’m not the only one shouting into the bot-infatuated void. Instagram account @pressrelease.sus is an amusing resource that highlights some dubiously em-dashed press releases (though I’m not convinced they’re ALL AI, but they are all sus). More seriously, in his excellent “Let’s Talk About Artificial Intelligence Art English” for Mousse Magazine, Andrew Berardini explores the rise of AI in art writing and launches an impassioned plea for maintaining the human voice. To him, AI represents “a betrayal of art’s essential purpose.” He continues, “If art exists to help us understand what it means to be human, then the language around art must itself be recognizably human: flawed, searching, alive with the particular consciousness that created it.”

Berardini echoes many of my own thoughts about AI art writing, except I’m much more cynical about the sorry state of art writing even without AI. Most art writing is so agonizingly boring and gratingly vacuous that it sounds as if it were written by a bot already, a mishmash of artspeak terms, like liminality, materiality, “space,” the gaze, etc., to avoid having to state a position on, well, anything. Previously, before having to face the sheer horror of abysmal AI art writing, I brushed off the influx of AI, presuming that if art writing is already mostly terrible, then who cares if it gets replaced? If your writing is so dull that it can be mimicked by a large language model that simply guesses the next logical word rather than writes from fully formed thoughts, that’s YOUR fault, not the tool.

What I didn’t account for is how goddamn atrocious the AI art writing would BE, particularly in gallery press releases. Some of these press releases are so filled with unnecessary em-dashes and loopy phrasing that it’s clear nobody even bothered to do a brief edit, which is a testament to how much contempt they really have for their audience. In Berardini’s essay, he generates his own example of a bad gallery exhibition press release because “I didn’t want to shame anyone in particular.”

Well, I do. The only way we stop this AI art writing invasion is to mock it publicly, loudly, ruthlessly, and relentlessly. So, I’m resurrecting my Drag Them to Filth series, which previously ripped apart terrible art criticism and cultural journalism. But at least that was by humans. Now, years later, the bigger danger and embarrassment is this AI slop, as evidenced by the absolute worst AI-generated press release I’ve ever read. The press release comes courtesy of 303 Gallery’s Hans-Peter Feldmann exhibition and is so hilariously, nearly feloniously bad that I found myself cackling about it all last week. Now, I should say that there is a (very small) chance a human wrote this, but if so, get someone else to do it next time because…Jesus… So, let’s break it down:

After the rote “please to present…” rigamarole, the baffling writing starts slow but quickly accelerates:

“Düsseldorf-based Hans-Peter Feldmann was a passionate collector of images and ephemera, an original thinker and one of the first conceptual artists. Feldmann focused on photography and artist books since the late sixties. Often collaging found images from magazines, postcards, books, advertisements, and stamps, Feldmann used familiar motifs to explore the boundaries between art and everyday life.”

Vague but passable.

“Unmade beds, car radios – Feldmann focalized unattended moments for a contemplative narrative based on being rather than acting.”

Hm….?! Have you ever used “focalized” in a sentence? Me neither! What is the difference between being and acting with an unmade bed? How about a car radio?

“Influenced by the Dadaist, the Situationists, Fluxus and Vienna Actionists, Feldmann viewed art as an impression rather than the object. Feldmann’s practice was defined not by materiality but by ideas.”

Profound! What does it mean?

“On view are several works from Feldmann’s Time series which began in 1970, which use analogue film rolls to shoot continuously the same place, object, or person. Modest subjects and trivial themes are chronicled frame by frame, recording the passage of time. His methods focused on the sequence rather than a single frame; it was through repetition that ideas became more conscious and deeper insight could be gleaned.”

Trivial themes! I’m gleaning right now!

“His work Beine consists of 30 photographs of women’s legs – where one photograph of a woman’s legs may simply be provocative, 30 photographs pinned together becomes something more.”

Are women’s legs provocative? In 2025? Do you want to take a look at my sexy ankles on OnlyFans? And what is this something MORE?

“In Alle Kleider einer Frau, 70 pieces of woman’s garment are displayed one photo at a time, the woman hidden, the artist emotionally distant, opening up space for the viewer’s interpretation.”

Thank god we have space for interpretation! And why is Hans-Peter so emotionally distant? Come back, Hans!

Now, for one of my favorite lines:

“In 2007, Feldmann returned to painting without assuming the traditional role of the painter.”

How innovative! A painter…but not a painter! The rest of the press release is, by comparison, pretty unremarkable, up until the last line, which is so transcendently dumb I found myself compulsively reciting it in the shower like a prayer:

“Feldmann believed art isn’t confined to a gallery or museum but woven into everyday life – his works are a reflection on looking, an ability that every viewer holds.”

Yes, we do all have the ability to look. Or most of us. Kind of leaving out people who are blind here. Ableist! And for the rest of us, how moving! Though…isn’t most visual art by definition “a reflection on looking”? Isn’t that the thing? I love that nobody at 303 Gallery decided…maybe we snip that last bit as it sounds completely ridiculous. I’m so glad they didn’t. Thanks for the laughs!