Throughout my years of writing about art, I’ve, of course, encountered many, many terrible press releases, filled with impenetrable jargon, effusive yet meaningless artspeak, and so much liminality, relationality, materiality, and space. I’ve rolled my eyes. I’ve sighed. I’ve cackled. Yet, it wasn’t until recently, with the overwhelming AI sloppification of NYC art galleries’ written materials, that a heinous press release completely ruined the show for me. The tipping point was Slip House’s just-closed group show Vestige, which brought together Medieval miniatures with contemporary artists drawing on the legacy of that gorgeous gilded religiosity. Though the exhibition itself was overpowered by the sheer terror of the gallery’s perilous third-floor spiral staircase, which made me break out into a cold flop sweat, I considered attempting to race out a review before its closing so I could gush over Ben Cowan’s holy blessing fingers (just a digit or two away from the sacred shocker); Fabiola Jean-Louis’s golden invocation of Haitian Vodou buried within Catholic imagery; Jennifer Carvalho’s sensuous fixation on Saint Catherine’s patterned robe; and Abbey Muza’s bare-there vulvic side wound tapestry.

It wasn’t until the next morning, when I rescued the crumpled paper from my tote, that I read the press release. And was instantly turned off. Why? Well, this AI-generated PR perseveration is worth quoting at length so you, too, can experience the bafflement of bot bizarreries:

“Images rarely stay still. They travel—across time, across material, across hands—carrying fragments with them, shedding others, returning altered. Vestige brings together Abbey Muza, Jennifer Carvalho, Ben Cowan, and Fabiola Jean-Louis around this sense of the image as something in motion: not fixed in the past, but continuously re-formed in the present.”

Is shedding fragments like shedding skin cells?

“At its center, a new body of work by Abbey Muza unfolds as a kind of woven language. Drawing from medieval Apocalypse tapestries and the symbolic systems of midcentury artist Jean Lurçat—alongside contemporary strategies of encoding and signaling—their dye-printed textiles hold images as they pass between clarity and abstraction. Ferns, fire, water, stars: forms that have carried meaning across centuries reappear here as signals—of renewal, warning, environmental precarity—while also remaining open, unstable, and in flux.”

Wait…you mean these are open…unstable…AND in flux?! And I didn’t realize signaling was a contemporary strategy!

“Muza’s process is both deliberate and responsive. Imagery is painted onto silk threads before weaving, so that as the cloth is formed, the image shifts—softening, misaligning, or dissolving into structure. Woven in doublecloth, each textile holds two layers at once, binding surface and ground together. The works are divided into quiet, stacked frames that isolate and re-sequence fragments, as if time itself were being edited into parts. Alongside them, a series of drawings returns to these forms—reframing, tracing, or clarifying what the woven image leaves unresolved. Meaning moves between them, never fully settling.”

 

What does it mean to soften, misalign, or dissolve into structure? What are quiet frames? What is a loud frame?

“This sense of the image as something held in suspension continues throughout the exhibition.”

Yeah…I bet.

“Jennifer Carvalho’s paintings gather references that feel half-remembered—Antiquity, Renaissance composition, cinematic space—layered into shallow, luminous surfaces where forms emerge and recede at once.”

They’re paintings. The forms have already emerged.

I assume you get the point, so let’s skip to the Medieval works:

“Their surfaces—worked with gold, pigment, and script—reward sustained looking, unfolding slowly rather than yielding all at once.”

“Placed in proximity, these works do not illustrate a lineage so much as share a condition: images made to endure, to carry meaning across time, and to be continually re-encountered. They ask for attention not as relics, but as active participants in the present.”

“Placed in proximity”…yes, this is an art exhibition. And condition?! Maybe if I stare long enough, I can see the condition their condition is in!

And my favorite:

“Vestige lingers in what remains: fragments, symbols, impressions that persist and shift. Not the image as it once was, but as it continues—threaded through time, held in material, and always, quietly, becoming.”

I, too, am quietly becoming.

Ugh. I don’t want to seem like I’m only picking on Slip House. Vestige was just my last straw, but it is not the first show that has been ruined for me by the slop ick. At least I read the press release after I saw the show, unlike PPOW Gallery’s description of John Kelly’s A Friend Gave Me A Book, which tanked the exhibition for me even before I elbowed a bunch of tourists buying knock-off wallets to get in the door. It was this paragraph that did it:

“However, Kelly’s transposition of tragedy into visual memoir is importantly not proscriptive. Rather than use the personal to address the political, the artist instead offers multiple inroads for dealing with feelings of isolation and sense of self, in which reality becomes a palimpsest of interior monologue, human interaction, cultural allusion, and externalization. In the artist’s words, “My works flow out of a desire to embody a particular challenge or event, whether from imagination, actual persons’ reality, or personal history: life imagined, life considered, life experienced.” In this way, Kelly uses art to invite his audience into his illustrated world, not as witnesses, but as participants in the quest for consequence and acceptance.”

1. I don’t even think “proscriptive” is the right word there. 2. Reality as a palimpsest of interior monologue, human interaction, cultural allusion, and externalization? Makes sense to me! And 3. I’m happy to be a participant in the quest for consequence! The same pre-show PR disgust occurred with Ortuzar’s current Peter Hujar exhibition, in which the large-language model even went so far as to explain what a photography show generally looks like:

“Matted and hung inches apart, the photographs are sequenced so that images from the same genre rarely appear consecutively…This arrangement emphasizes the singularity of each person, place, or animal, inviting viewers to move in and out of the grid as visual and emotional connections emerge, dissolve, and reform. As Hujar once noted, “I photograph those who push themselves to any extreme. That’s what interests me, and people who cling to the freedom to be themselves.” Rather than comparing his subjects with each other, he was determined to see the singularity in each, an aim the exhibition supports.” 

Wait…the singularity?! I thought that was some Einstein shit. When I Google singularity, I get this: “technological singularity is a predicted future point where exponential advancements in AI create superintelligence, surpassing human intelligence and causing unpredictable, runaway technological growth.” Is THIS the aim the exhibition supports, or is it the aim of the AI chatbot that generated this?

I know what you’re thinking: Emily, stop being paranoid. How do you know this is AI rather than a bored gallery assistant? Well, AI-slop writing is not that hard to discern. Sure, amateurs will point out the em-dashes, which gugs me, given I overused em-dashes way before Claude was even a sparkle in Dario Amodei’s eye. More discerning readers will note certain red flag off-kilter sentence formations like “not_____, but_____,” as in Gerne en Regalia’s current “Companion gestures toward a world expanded not through clarity, but through intimacy, pause, and the generative potential of not yet knowing,” or the similar, “less a _____ than a/more a _______,” like Fierman’s “color functions less as a means of depiction than as a condition…” and Chapter NY’s “less a depiction than a condition of an image…” So many conditions!

As seen in the epidemic of conditions, I’ve noticed a series of odd trends emerging and reemerging in AI exhibition press releases. All across NYC, fluid, unstable, in-flux images, subjects, and their meanings are dissolving, slipping and sliding, and reforming. Chapter NY’s Mary Stephenson show features “images pausing, replaying, zooming, and moving in rhythm with the passage of time.” At Company Gallery, Stefania Batoeva’s figures are so slippery that they aren’t even figures:

“In Batoeva’s paintings, figures don’t hold. They loosen, slip into context, and surface, hovering somewhere halfway in abstraction. In fact, they don’t quite arrive as figures at all. Closer to persona than to form, they remain unstable, elusive, and ever-evolving. Batoeva’s figures are shapeshifters, changing the story and reassembling themselves as they move through dense, atmospheric fields. They leave behind only partial phrases.”

Not even 93-year-old Joan Semmel’s excellent show at Alexander Gray Associates, with its aggressively wacko color scheme and grand landscapes of aging female flesh (my favorite, a lumpy tummy and belly button), escapes this treatment. As observed on a wall text: “Color moves across the artist’s flesh in sustained passages while bodily contours soften and re-emerge.”

THEY’VE ALREADY EMERGED!

What is so remarkable about this samey-samey language is not only that galleries have such a low regard for their own visitors that they don’t care if they all say the exact same thing, but that AI models are generating this schlock at all. Large-language models gobble up all of our previous human writing and spew it back at us, which means that AI art press releases should be a compendium of previous art writing. Yet, did anyone ever write like this? I don’t recall! In fact, I’m filled with existential terror that I may have published something like this at any point in my career. Was it ME?! Am I responsible? I can’t imagine! My theory is that AI has now reached sentience—the singularity, if you will—and this is how the machines understand art. Doesn’t it SOUND exactly like how a computer would analyze a painting?

And because I can’t restrain myself, I have to pause to highlight the conclusions of two of the just-mentioned press releases because they’re completely insane and make me feel like I’m having a stroke (is that toast?). First up, Chapter NY:

“If painting carries a thought from one side to another, from sensation to memory and from self to world, it does so not by constructing a bridge, but by entering the water and finding a way through. Not a performance, but a dance.”

Dipping into the water and paddling my way through to Company Gallery’s berserk articulation of a rumpled Ikea bed plunked in the middle of a painting show:

“At the center of the exhibition, a spatial intervention lures for a pause: a king-sized bed complete with pillows and bed linens is positioned against a dividing wall below one of Batoeva’s paintings. To be here, not quite outside the image and not fully inside it either, the viewer gets caught in a cycle of looking that turns back on itself. The space quiets and folds into the covers, becoming less of a room and more of a state of being. Between waking and dreaming, the installation reads like a soft break in the tempo, an opening into one of Batoeva’s parallel worlds. It makes room for rest and vulnerability, if only briefly, against a world that won’t.”

This is what that looked like:

As amusing as it is to gape and chuckle at the absurdity of this cheapo mattress, barebones bedframe, and abysmal depression grey sheets as a state of being, the rampant scourge of AI-slop press releases is the most soul-deadening development in the art world yet (and that’s saying something). It really toasts my buns for several reasons. First, as impenetrable and oblique as artspeak has always been, exhibition descriptions and wall text exist for a reason. They’re not useless exercises to see how many words you can use without creating any meaning whatsoever. Press releases give context to the exhibition—who the artist is, their relevant background, and some of the concepts behind the work on view. These are entries into exhibitions for people who need a bit more detail—or, too often, art writers who need an existing text to give them ideas. What did I learn from any of the aforementioned press releases other than imagining a free-flowing, ever-evolving, always-in-transition stream of imagery that flows through the art world like eye floaters on a sunny day? You might as well not have a press release. Fuck it.

Secondly, one would think that of any industry, the art industry would maintain its respect for human insight, human creativity, and human interpretation of other humans’ work. Ok. I know. I know. I laughed too. But really, turning art analysis and contextualization over to AI exposes the art industry for what it really is—a soulless, vapid commercial enterprise that is no different than Amazon producing AI-slop podcasts to sell diaper cream. No wonder art sales suck. It’s probably more fun to gamble on whether a ship makes it through the Strait of Hormuz on Kalshi.

Then, of course, there are the ever-present environmental concerns, the coming-to-an-inbox-near-you AI-driven mass lay-off economic calamity, and the windfall profits for a handful of tech billionaires who want to spy on their fellow citizens while lecturing on the Antichrist and buying doomsday properties in New Zealand in case the masses get restless. Given that so much of the art industry (for some reason) seems to see itself on the right side of history, one would, again, assume someone at these spaces might think twice about participating in our own destruction at the behest of a Silicon Valley freakazoid like glassy-eyed Sam Altman. Or they might consider the poor people in rural America who are now being driven insane by a constant humming emanating from their new data center neighbor, because a gallerist in Tribeca couldn’t come up with a way to describe an abstract painting.

That lack of imagination gets to the final reason AI-slop art writing grinds my gears. I’m insulted personally as a writer. I know some will argue that galleries are just overworked, so they’re trying for the easiest, quickest option. Yet, the easiest, quickest option would be to commission a writer to do the job, which would be helpful since every other soul-deadening industry is giving up on human writers, too. Or just have the overeducated and underpaid gallery assistant do it. When I worked at Sikkema Jenkins years and years ago, writing the press releases was by far the most interesting part of the gig, other than preventing the plebs from using the gallery’s private bathroom. Now, I’m aware that perhaps these galleries did assign the job to a blasé intern or gallery assistant who had Gemini whip up this garbage so they could go back to doomscrolling. Yet, someone else had to approve of these literary atrocities, so that’s not a worthy explanation.

Part of me wonders if this heavy influx of obvious AI is evidence of a lemming-like groupthink (previously visible in the art world’s strike earlier this year) rampaging through the art world with AI usage as some bizarre sort of status symbol, similar to corporations aggressively pushing AI usage on employees so they can train their permanent replacement. It’s probably no mistake that when Artnet and Artsy merged late last month, their new owner used the opportunity for a masturbatory wax poetic about artificial intelligence: “We are at a pivotal moment in which data and technology, including artificial intelligence, are transforming every industry…The art world needs companies that can harness these powerful forces for its benefit.”

AI-generated Wet Paint gossip column coming to an inbox near you!

What’s the solution? Certainly, these galleries aren’t going to quit. Last year, I returned to my Drag Them to Filth series to mock appalling AI press releases, starting with 303 Gallery. Yet, the oozing AI infection is getting so rampant that it’s impossible to pick a worst out of the bunch. While I might reboot the series now and again for a particular rotter, there are so many that I could easily only cover dreadful AI art writing on this site. So, I feel forced to take a more extreme stance: I refuse to review any exhibition with AI-slop press releases from here on out. Yes, I realize this risks not writing about art much. That’s okay—I have other interests. There are also enough galleries and, mercifully, still most, if not all, museums that haven’t gone full bot yet (Case in point: MoMA PS1’s simple labels for Greater New York were a welcome human relief). I also realize that revoking a less-than-coveted Filthy Dreams review generally will not matter, but fuck it, SOMEONE has to have standards around here!

How will I know what press releases are AI slop? Trust me, I just do. In a bind, I could use one of the plethora of AI detector tools that desperate professors employ to see if their students are generating their papers (they all are). And those detectors can be a delight, like identifying the first two paragraphs of Andrew Edlin Gallery’s previous group show, Afterlife, as 100% AI (Nailed your ass!). But, in general, these tools are not all that intuitive. Luckily, I don’t need them. I can sniff out AI slop like a literary bomb-sniffing dog, or more accurately, a cadaver dog. I can catch a whiff of that dead language no matter how heavily a publicist attempted to edit it, like David Zwirner’s upcoming Lisa Yuskavage show. While sure, their team did a better job of massaging the AI slop, my spidey-slop-senses tingled at the description of the show consisting of “works that insist on being looked at.” Yes, it’s an art exhibition. Looking is kind of the point.

And that’s exactly what I’m going to do. I’ll look. But I’m not reviewing.