There’s a specific, dizzying nausea that hits when a painting you bled over for six months gets twelve likes, while a video of someone pouring glitter on a pizza slice racks up 2.3 million. It’s not jealousy—well, not just jealousy. It’s the creeping realization that the conversation you thought you were having, the one between you, your materials, and maybe a few hundred curious strangers, has been hijacked. The new referee is a stack of code that can’t tell a Rothko from a rug.
We’ve spent centuries bickering about what art is. Beauty? Truth? A urinal with a fake signature? Now the answer is getting simpler and dumber: art is whatever the feed lets through. The algorithm has become the final critic, and its taste is, to put it generously, chaotic. It rewards the legible, the repeatable, the instantly gratifying. It punishes the slow, the ambiguous, the thing that asks you to sit with discomfort. This isn’t a fresh complaint, but the mechanics are worth dissecting, because they’re reshaping not just what gets seen, but what gets made.
The Tyranny of the Thumb-Stopping Moment
Every platform’s algorithm runs on engagement signals: likes, shares, comments, and—most tellingly—dwell time. The longer your thumb hovers, the more the machine thinks you’re hooked. For artists, this is a brutal design brief. Your work has to function as a thumbnail, a split-second seduction in a sea of distraction. A Caravaggio, with its deep shadows and slow-burn drama, would bomb on Instagram. You’d scroll right past Supper at Emmaus because it doesn’t pop on a backlit screen at two inches wide. But a neon sign that says “FEELINGS” in cursive? That’s algorithmic catnip.
This isn’t just about subject matter. It’s about formal qualities. High contrast, saturated colors, symmetrical compositions, and human faces—especially with direct eye contact—consistently outperform subtler work. The algorithm has become a patron with very specific, philistine tastes. It wants art that behaves like content: easily digestible, emotionally legible, and formatted for rapid consumption. A Mark Rothko canvas, meant to be experienced in a quiet room at human scale, becomes a muddy brown rectangle on a phone. A Jeff Koons balloon dog, however, looks exactly like what it is, no matter the screen size. The medium isn’t just the message; it’s the curator, the critic, and the crowd.
This flattens the spectrum of artistic expression into a narrow band of what I call “scroll-stopper formalism.” You see it in the rise of hyper-detailed, photorealistic pencil drawings that are technically astonishing but conceptually vacant. You see it in the endless iterations of the same aesthetic: the pastel-colored flat lays, the glitter-drenched lips, the acrylic pours that look like planetary nebulae. These works are engineered for virality. They’re the visual equivalent of a pop hook—immediate, catchy, and forgotten in three seconds. The algorithm doesn’t care if you remember it; it only cares that you stopped.
The Death of the Difficult Object
What gets lost in this feed-first ecosystem is the art of friction. I’m talking about the work that resists easy consumption: performance art that unfolds over hours, installations that require physical presence, paintings that reveal themselves slowly through texture and scale. These are experiences that cannot be compressed into a 1080×1080 pixel square without losing their essential quality. And because they can’t be compressed, they effectively cease to exist in the digital public square.
Consider the fate of time-based media. A 20-minute video essay on the semiotics of surveillance might be a masterpiece, but it’s dead on arrival on TikTok, where the ideal video length is somewhere between 15 and 60 seconds. The platform’s entire architecture is hostile to duration. Even on YouTube, where longer formats survive, the algorithm punishes anything that doesn’t maintain a high “retention rate.” This means creators are incentivized to front-load their videos with hooks, to avoid silence, to cut every pause. The result is a kind of breathless, overcaffeinated style that leaves no room for contemplation. Art that asks you to be bored, even for a moment, is art that asks to be ignored.
The Aesthetic Echo Chamber
Algorithms don’t just select for certain visual traits; they create feedback loops that homogenize taste. When an artist sees that a particular style or subject gets traction, they’re incentivized to produce more of it. The algorithm then shows that work to people who engaged with similar content, who then engage with it, reinforcing the signal. Over time, this carves deep ruts in the cultural landscape. Entire genres emerge not from artistic movements or critical discourse, but from data patterns.
Take the phenomenon of “Instagram poetry.” Rupi Kaur’s spare, illustrated verses are perfectly adapted to the platform: they’re short, they’re emotionally direct, they look beautiful in a square format. Her success is genuine, but it also spawned a thousand imitators who mistook the format for the substance. The algorithm rewarded the form—the line breaks, the lowercase letters, the simple sketches—and so the form proliferated, regardless of whether the content had anything new to say. This isn’t Kaur’s fault. It’s the logical endpoint of a system that optimizes for replicability. The algorithm loves a template.
This templatization extends to visual art. Search for “contemporary art” on Pinterest, and you’ll find a strangely uniform vision: resin-coated everything, geometric line drawings of faces, watercolor splashes behind motivational quotes. These aren’t trends in the traditional sense, driven by curators or critics or even artists talking to each other. They’re trends driven by the platform’s recommendation engine, which notices that people who like one resin pour tend to like other resin pours, and so it shows them more resin pours, and so more artists make resin pours, and the cycle tightens until the entire category becomes a hall of mirrors.
When the Market Becomes the Medium
Historically, the art market has always been a strange, often corrupt mediator between artists and audiences. But the social media market operates differently. It’s not just about money; it’s about visibility as currency. An artist with 100,000 followers can bypass traditional gatekeepers—galleries, critics, grant committees—and go straight to an audience. That’s genuinely liberating in some ways. It has democratized access, allowing artists from marginalized communities to find audiences that the old guard ignored.
But this new gatekeeper is, in many ways, more insidious than the old one. A gallerist might have bad taste, but they’re at least a human with a point of view. The algorithm has no point of view. It has metrics. It doesn’t care about your identity or your politics except insofar as they generate engagement. It will happily amplify a queer artist of color one day and a fascist meme the next, as long as the numbers are good. This isn’t neutrality; it’s a kind of nihilistic optimization that treats all content as equivalent fuel for the attention engine.
And then there’s the question of value. When visibility is the primary currency, artists start to internalize the algorithm’s logic. They begin to ask not “What do I want to make?” but “What will perform well?” This is a subtle but profound shift. It turns the studio into a focus group, the sketchbook into an A/B testing ground. The work that emerges might be popular, but it’s often hollow, because it was never really made for anyone—it was made for the machine.
Resistance Is Possible (But It’s Weird)
So what’s an artist to do? Quit the internet? That’s one option, but it’s a privileged one. For many, social media is the only viable way to reach an audience, sell work, and build a career. The more interesting move is to hack the system, to use the algorithm’s logic against itself. Some artists are doing this by creating work that explicitly critiques the platforms it’s shown on—posting glitch art that corrupts the smooth feed, or performance documentation that refuses to explain itself. Others are building intentional communities on slower platforms like newsletters or Discord servers, where the pace of interaction allows for deeper engagement.
There’s also a growing movement of “algorithmic disobedience.” This might mean posting work that’s deliberately unphotogenic, or using the platform to direct people toward physical experiences. I’ve seen artists post a single, cryptic detail of a larger installation with a note: “You have to be there.” It’s a refusal to translate the work into feed-friendly terms, a demand that the audience meet the art on its own ground. This is risky, of course. It might mean fewer likes, fewer followers, less “reach.” But it also might mean a more honest relationship with the people who do show up.
The point isn’t to demonize social media. It’s a tool, and like any tool, it shapes the hand that wields it. The danger is when we forget that the tool has a logic of its own, and that logic is not neutral. Every time you post, you’re making a choice: to feed the machine what it wants, or to feed it something it doesn’t know how to digest. The latter is harder. It might not “work.” But it’s the only way to keep the machine from deciding what art is before we do.
Frequently Asked Questions
Does social media actually change what kind of art gets made?
Absolutely. Artists are practical creatures. When they see that certain styles, formats, or subjects consistently get more engagement, they naturally gravitate toward those. This isn’t always a conscious sellout—it’s often a survival strategy. The result is a flattening of aesthetic diversity, as the market rewards work that fits the algorithmic mold. You can trace direct lines from the rise of Instagram to the explosion of certain visual trends, like flat lays, neon signage, and hyper-saturated portraiture.
Isn’t this just the same as artists always following trends?
Not quite. Traditional art trends were shaped by human networks: critics, curators, collectors, and artists in conversation. Those networks had biases, but they also had memory and context. An algorithm has no memory beyond its training data and no context beyond the immediate engagement metrics. It can’t understand irony, historical reference, or subtle provocation. It just sees patterns. So the trends it creates are shallower and more volatile, because they’re not rooted in any shared cultural conversation—they’re rooted in what made people stop scrolling for 0.2 seconds last Tuesday.
Can “difficult” art survive on social media?
It can, but it requires a different strategy. Difficult art—work that’s slow, ambiguous, or demanding—can’t compete on the algorithm’s terms. Instead, it needs to build a different kind of relationship with its audience. This might mean using social media as a pointer rather than a gallery: posting process shots, studio visits, or contextual essays that draw people toward the actual work. Or it might mean abandoning the major platforms altogether for spaces that allow for deeper engagement, like newsletters, podcasts, or in-person events. The key is to stop expecting the algorithm to do the work of a curator. It won’t, and it can’t.
What’s the alternative to algorithm-driven art?
The alternative is art driven by human attention—slow, deliberate, and often inconvenient. It’s art that asks you to put down your phone and be in a room with something. It’s art that’s made for a specific community rather than a generic “audience.” It’s art that might only reach a hundred people, but reaches them deeply. This doesn’t mean rejecting technology. It means being intentional about which technologies you use and why. A zine passed hand to hand, a performance documented in a single photograph, a painting that takes a year to finish—these are all technological choices, and they all resist the feed’s demand for speed and scale.