You think you choose the art you love. A painting grabs you in a gallery. A song stops you mid-scroll. A film leaves you gutted in a dark theater. But let’s be real: that first encounter is rarely a choice anymore. It’s a feed. And the feed has its own logic, its own rhythm, its own preferences—none of which are yours. The algorithm is the most powerful, least visible curator of the twenty-first century. It doesn’t hang canvases on walls. It hangs pixels in front of eyeballs, and it gets paid when you linger.
This isn’t another eulogy for taste. Taste is alive. But it’s being nudged, trained, and sometimes flat-out manufactured by systems built to maximize engagement. The real question isn’t whether the algorithm influences what we see. It’s how deeply it has reshaped what gets made in the first place.
The Engagement Engine Doesn’t Care If It’s Beautiful
Let’s not mince words. The algorithm’s job isn’t to surface the sublime, the challenging, or the genuinely strange. Its job is to keep you glued. Instagram, TikTok, X, YouTube—they all optimize for dwell time, shares, comments. The metric isn’t aesthetic merit. It’s frictionlessness. A visual that’s instantly legible, emotionally unambiguous, and loud will trounce a quieter, weirder work almost every time. This isn’t a conspiracy. It’s just math.
Think about the fate of the difficult image. A photograph that wants you to sit with discomfort. A painting that refuses to resolve. A short film with no clear climax. These works demand time, and time is the one thing the feed can’t afford to give you. The algorithm is a hyperactive docent, tapping your shoulder every three seconds: “Still entertained? No? Next room.”
Artists have always worked against constraints. Patrons wanted flattery. The Church wanted piety. The market wanted scarcity. But the algorithmic constraint is weirder because it’s invisible and absolute. It doesn’t announce itself. It just withholds distribution. An artist might never know why their work flatlined. They just see the numbers and, consciously or not, start adjusting.
The Rise of the Optimized Aesthetic
Scroll any visual-forward platform long enough and you’ll spot patterns that aren’t just trends. They’re structural tics, baked in by the feed’s demands. High-contrast palettes. Central compositions that read at thumbnail size. Faces, always faces, preferably making eye contact or telegraphing one unmistakable emotion. Text overlays that explain the joke before you can miss it. The work comes pre-digested.
I’m not dragging the artists who make this stuff. Plenty of them are brilliant, and they’re doing what artists have always done: responding to the conditions of their visibility. But let’s be honest about what those conditions reward. The algorithm loves legibility. It loves speed. It loves sentimentality you can process in under a second. It has no patience for ambiguity, slowness, or the kind of beauty that only reveals itself after repeated looking.
What we get is a global vernacular—a visual language native to no particular culture but perfectly adapted to the feed. You see it in illustration, photography, graphic design, even in how painters compose canvases for Instagram. The work is often gorgeous, technically dazzling, emotionally immediate. It is also, increasingly, the same.
The Flattening of Weirdness
Here’s where it gets interesting, and a little sad. The internet was supposed to be a paradise for the niche, the eccentric, the gloriously unmarketable. And for a while, it was. Tumblr gave us communities organized around aesthetic fragments so specific they barely had names. Early YouTube was a fever dream of personal obsession. But as platforms consolidated and algorithms tightened their grip, the weird got smoothed out.
Why? Because weirdness doesn’t scale. A deeply peculiar video might delight ten thousand people and utterly baffle ten million. The algorithm, tasked with maximizing total engagement, will favor the thing that mildly delights ten million over the thing that electrifies ten thousand. The result is a kind of cultural averaging. The edges get sanded off. The strange becomes content-adjacent. The genuinely transgressive gets buried under a mountain of mildly spicy takes.
This has real consequences for what gets called art. When the main distribution channel for visual culture rewards broad, instant appeal, the definition of “good art” starts to warp. It becomes synonymous with “highly shareable art.” The work that spreads is the work that confirms what people already feel, not the work that complicates or unsettles. We’re building a canon on the lowest common denominator of emotional response.
The New Gatekeepers Wear Hoodies, Not Tweed
Art history is a parade of gatekeepers. The academy. The salon. The critic. The gallerist. The collector. Each had a taste, a bias, a set of interests that shaped what was seen and what was forgotten. But those gatekeepers were at least human. You could argue with them. Seduce them. Scandalize them. If you were clever and lucky, you could route around them.
The algorithm is different. It’s not a person with discernible taste. It’s a bundle of optimization functions wrapped in machine learning models, constantly updating based on what works. You can’t argue with it. You can’t write it a manifesto. You can only feed it data and hope. The engineers building these systems aren’t art critics. They’re not trained in visual culture. They’re solving problems of latency, retention, and ad load. And yet their decisions—what counts as a view, how long a video must play to register, whether a like weighs more than a share—are aesthetic decisions. They’re shaping the sensory environment of billions of people.
This is a new kind of power, and it’s almost entirely unaccountable. No one elected the recommendation algorithm. No one debates its values in public. It just sits there, silently reordering the world’s images according to a logic no single human fully understands.
Resistance Is Possible, but It Looks Different Now
So what’s an artist supposed to do? Submit? Withdraw? Game the system? The most interesting responses I’ve seen do none of these cleanly. They treat the algorithm as a material to be worked, like clay or light. Some artists create work that’s deliberately illegible to the feed—images too dark, too slow, too strange to perform well—and then document the algorithm’s failure to digest them. The rejection becomes part of the piece. Others build their practice entirely outside the platforms, using newsletters, print, or in-person gatherings, accepting a smaller audience in exchange for a more honest relationship.
There’s also a growing movement of artists who hack the logic of virality. They make work that looks, at first glance, like optimized content—bright colors, bold text, familiar formats—but delivers something unsettling on closer inspection. It’s a kind of algorithmic judo, using the platform’s own weight against it. The work gets distributed because it passes the engagement checks, but it leaves the viewer slightly queasy, slightly less sure of what they just consumed.
None of these strategies are perfect. They all involve compromise. But they suggest that artists aren’t passive victims of the feed. They’re learning to see the algorithm as a condition, not a destiny—a weather system to be navigated, not a god to be obeyed.
What Happens to the Audience?
If the algorithm shapes what artists make, it also shapes what audiences expect. We’re training a generation of viewers to demand instant gratification from images. Scrolling is a form of looking that’s rapid, evaluative, and binary: yes or no, stop or keep moving. This is very different from the kind of looking that happens in a museum, where you might stand before a single work for minutes, letting it unfold. The feed doesn’t teach patience. It teaches appetite.
And appetite, unchecked, becomes a kind of blindness. When you’re used to images that announce their meaning loudly and immediately, you lose the ability to read quieter work. You start to find ambiguity annoying rather than intriguing. You mistake slowness for failure. The algorithm doesn’t just shape what art gets seen; it shapes the sensorium of the viewer, the very capacity to perceive.
This is the deeper loss. Not that some good artists go undiscovered—that’s always been true. But that the audience itself is being trained out of the skills required to encounter difficult art. When we finally do stand before something that asks for time, for patience, for a willingness to be confused, we may find we no longer know how to give it.
Frequently Asked Questions
Does the algorithm actually change what kind of art gets made?
Yes, and in more ways than you might think. Artists who rely on social media for visibility and income—which is most emerging artists today—inevitably internalize what performs well. This can mean brighter colors, simpler compositions, more explicit emotional cues, and formats optimized for tiny screens. The effect isn’t always conscious. It seeps in through the feedback loop of likes, shares, and comments. Over time, the feed’s preferences become the artist’s intuition.
Is this really different from old gatekeepers like galleries and critics?
In one sense, no. Every distribution system has biases. The French Academy favored history painting over still life. Mid-century critics championed abstraction over realism. But those biases were articulated. You could read Clement Greenberg and know what he valued. You could argue with him. The algorithm’s biases are embedded in code and constantly shifting, and they’re optimized for engagement, not for any coherent aesthetic philosophy. That’s a meaningful difference.
Can good art still break through?
Absolutely. The feed isn’t a totalizing force. Strange, slow, difficult work still finds audiences—sometimes because it goes viral against all odds, sometimes because it builds a small, devoted following that the algorithm can’t see. But the path is narrower than it used to be, and the pressures are real. The danger isn’t that good art disappears. The danger is that we stop being able to tell the difference between what is good and what is merely optimized.
What can viewers do to resist the flattening?
Seek out work in spaces that aren’t algorithmically sorted. Go to galleries. Subscribe to independent art publications. Follow artists to their personal websites and newsletters. Buy prints directly from creators. And when you do scroll, do it with awareness—notice what the feed is feeding you, and ask what it’s not. The algorithm is a curator with a conflict of interest. The best response is to become your own curator, deliberately, even if it takes more effort.
The feed isn’t going anywhere. But we can learn to see it for what it is: a tool, not a tastemaker. Art has survived popes, kings, and market crashes. It will survive the algorithm too. The question is what kind of art we will become capable of making—and seeing—along the way.