How the Camera Created Abstract Art — and What AI Means for Photography
In 1839, a man in Paris named Louis Daguerre announced a process that could fix a real-world image onto a polished silver plate using mercury vapor and a few minutes of sunlight. The painter Paul Delaroche, looking at one of the first daguerreotypes, is supposed to have said: “From today, painting is dead.”
He was wrong about painting. He was right about something else.
What died was painting's job. For four hundred years, painters had been the only people on Earth who could record a face, a battle, a coastline, or a wedding party. The camera ended that monopoly in a single afternoon. And what painters did over the seventy years that followed — the path from Monet's haystacks to Pollock's drip canvases — was not a stylistic accident. It was a forced migration. They had to find something a camera could not do.
I am writing this in 2026 as a working photographer.I have shot for editorial fashion and won awards I am proud of, and over the last three years I have watched my own profession start to live through what painting lived through after 1839.AI now produces flawless technical realism in seconds.
The thing photography has done well for almost two centuries — record what was in front of the lens — is no longer a profession-defining skill.We have to find what only photography can still do.
The good news is that we already have a map. The painters drew it for us a hundred and fifty years ago.
Every time a machine learns to copy a human craft, the humans who survive are the ones who notice what the machine cannot copy — and run toward it.
1839: The job painters lost
Before the camera, if a wealthy family wanted a likeness of their daughter, they paid a portraitist for weeks of sittings. If a newspaper wanted a battlefield, they sent an engraver to sketch it. If a town wanted a record of its cathedral, they commissioned a landscape painter. Painting was infrastructure. It was how a civilization remembered itself.
In 1839 a daguerreotype could capture a sharper likeness of that daughter in five minutes. By the 1850s, wet-plate collodion brought the price down again. By the 1880s, dry plates made the camera portable enough to follow armies and weddings. Within fifty years, the bottom fell out of every market that paid painters to record what was visible.
Faced with this, painters had three choices. They could keep painting realism for the shrinking number of patrons who still preferred handwork. They could quit. Or they could find a reason to paint that had nothing to do with copying what the eye saw.
The third choice produced almost everything we now call modern art.
Monet: paint the light, not the thing
Impressionism — from object to perception
Claude Monet exhibits Impression, Sunrise. The harbor is barely there. What the painting captures is the orange of the sun bleeding through the morning fog. Critics use the word “impression” as an insult. Monet keeps it.
Monet's move was simple and brilliant: he stopped trying to paint things and started painting the way light changes things. Photography in his era was monochrome and fixed-exposure. It could record a haystack, but not how a haystack at dawn was a different color than a haystack at noon. Monet painted that haystack twenty-five times. The series was the point. He was painting time, not hay.
This is the first lesson the painters teach us: when the machine takes the surface, you go to the layer above the surface. You paint what the machine cannot record because the machine does not have a body, a memory, or a sense of duration.
Van Gogh: paint the feeling, not the field
Post-Impressionism — from accuracy to emotional truth
Vincent van Gogh paints a yellow wheatfield under a yellow sun, surrounded by yellow walls. The yellow is not the yellow of the wheat. It is the yellow of standing in that wheat with his particular mind on his particular afternoon.
Van Gogh is the second lesson, and it is harder than Monet's. Photography can record light, eventually, in color. What photography cannot record is what it felt like to be the one standing there. Van Gogh painted his nervous system. The brushstrokes are not decoration — they are the literal trace of his hand, his heart rate, his weather. A brushstroke is a fingerprint. A fingerprint cannot be photographed onto a separate canvas.
This is the second migration: when the machine can copy what was there, the human moves to what it felt like to be there. The artwork stops being a window onto a scene and starts being evidence of a person.
Picasso: paint how the mind sees, not how the eye sees
Cubism — from one viewpoint to all viewpoints at once
Pablo Picasso paints Les Demoiselles d'Avignon. Five women, but each face is wrong on purpose — profile and front view collapsed into one plane. With Georges Braque, he goes further. By 1910 a cubist portrait shows you the front of a man's head, the side of his head, and the shape of the room behind him, all on the same canvas at the same time.
The camera is locked into one fixed point of view. Click. The lens was here, the subject was there, this is what the geometry produced. Picasso noticed that this is not actually how human perception works. When you look at a friend, your eyes move, your head moves, your memory fills in the parts you cannot see. A real perception of a face is a composite. Cubism is not a worse drawing of a face. It is a more honest one.
The third lesson: when the machine can record one viewpoint perfectly, the human composes many viewpoints — or memory, or time, or a relationship — into a single image. The artwork starts depicting how the mind sees, not how a single eye sees.
Kandinsky: leave the subject behind entirely
Pure abstraction — from object to feeling, with no object in between
Wassily Kandinsky removes the subject. His Composition series uses color, line, and shape with nothing in front of them. He calls his paintings “compositions” on purpose, as if he were a composer. The painting is no longer a picture of a thing. It is a thing in itself.
Kandinsky is what happens when you take Van Gogh's and Picasso's logic to its conclusion. If the painting's job is to record a feeling rather than a place, why bother with a place at all? You can just record the feeling. A photograph cannot do this. A camera needs something in front of it. Pure abstraction is the first art form whose entire reason for being is that it cannot, by definition, be photographed.
Notice the trajectory. In 1839, painting and photography were rivals trying to do the same job. By 1913, they were not even in the same conversation. Painters had moved into territory the camera could not enter.
Pollock: the painting becomes the action
Action painting — from image to evidence of a body
Jackson Pollock lays an unstretched canvas on the floor of a barn and walks around it, dripping paint from a stick. The finished work is not a picture of anything. It is the record of where his arm went, how fast, for how long. The painting is a fossil of a performance.
By 1947, photography had become the universal record. Newspapers, magazines, family albums, military intelligence, advertising — the camera had eaten everything. Pollock's answer was the most un-photographable image possible: a painting whose entire content is the trace of a human body moving through a specific room over a specific hour. You can photograph a Pollock canvas, but you cannot photograph the Pollock. The painting is the trace, and the trace cannot be re-recorded.
This is the fourth and final lesson. When the machine has eaten the image entirely, the human moves into process. The artwork is no longer a noun. It is a verb that left a residue.
The pattern, in one paragraph
Every time a new machine learned to copy what painters were doing, painters did not stop painting. They moved one floor up. The camera could copy a face, so painters started painting light on a face. The camera could copy light, so painters started painting feeling. The camera could copy a single viewpoint, so painters started composing many viewpoints. The camera became the universal record, so painters made the painting itself a verb — the trace of a body in a room. At every step, the human went somewhere the machine could not follow yet. The work that survives is the work that was made on the floor above wherever the machine was standing.
Now: AI is the new camera, and photography is the new painting
I am going to say something that I think most working photographers already know but have not yet said out loud. The thing AI did to image-making between 2022 and 2026 is exactly what the camera did to image-making between 1839 and 1880. Faster, but the same shape.
Three years ago, “photorealistic” was a skill that took ten years of craft to acquire. Now any literate person can describe a scene in a sentence and receive a flawlessly exposed, flawlessly composed image of it. Skin tone, depth of field, color grading, golden hour — all of it, instantly, free.
In 1839, painting's monopoly was on realism. AI did not break photography's monopoly on realism — it broke photography's monopoly on realistic image-making. There is a difference. A photograph is a chemical or electronic recording of light that bounced off a real object. AI does not need the real object. The real object was the entire reason photography existed for 187 years.
So — like Monet in 1872 — we have to find a reason to keep walking out of the house with a camera that does not depend on the surface the machine has already taken.
What only photography can still do
The painters' map gives us four floors to climb to. Each one has a photographic version.
1. Witness — the “it actually happened” layer
A real photograph is evidence that a particular human body, at a particular instant, was at a particular place pointed in a particular direction.AI cannot supply this.AI can fabricate what looks like Aleppo in 2016 or a courtroom in 2025, but it cannot have been there.
As AI imagery becomes indistinguishable from reality, the value of being able to prove that an image is a real witness goes up, not down.Photojournalism, documentary, family archive, evidence — this layer becomes more valuable, not less.
2. Presence — the Van Gogh layer
The signature of this photographer's nervous system on this moment. How a photographer leans in. What they decide is the picture and what they let leave the frame. Why they wait. What they care about. AI can imitate a style; it cannot have the relationship a real photographer has with a real subject across years. Personal, repeated work — bodies of work — becomes the moat.
3. Composition of the unrepeatable — the Picasso layer
The decision to stand in the one spot, at the one second, with the one focal length, where geometry resolves into meaning. Cartier-Bresson called it the decisive moment. AI can synthesize a composed image, but it cannot stand in a moment that is happening for the first and only time. As AI eats studio realism, the value of being there for the unrepeatable second goes up.
4. Process — the Pollock layer
The making is the work. Long-form documentary projects, artist books, slow-built series, hand-printed darkroom work, alternative processes, photography that includes the artist's body and time as part of the medium. The Pollock move for photography is not a single picture; it is a process whose evidence includes the photographer.
And, like the painters who used cameras: use AI
The painters who survived the camera did not pretend the camera did not exist. Edgar Degas worked from photographs. Walter Sickert taught from them. Francis Bacon built his entire studio practice around them. David Hockney has spent the last fifty years experimenting with whatever new lens or scanner came out that year. The serious painters used the new tool while pursuing what only painting could do.
Photographers should do the same with AI. Use it to previsualize a shoot. Use it as a moodboard. Use it to scout a look before you book the location. Use it to test wardrobe, palette, and light direction before the model arrives. Use it the way you used to use Polaroids and reference scrapbooks. Then walk out the door with a camera and capture the moment that actually happened — the one that AI, by definition, cannot photograph because the moment had to occur.
The photographers who lose the next decade will be the ones who fight AI head-on for the same square inch of realism the painters lost in 1839. The photographers who win the next decade will be the ones who climb the floors the painters climbed: witness, presence, the unrepeatable instant, the process. AI is not the end of photography. AI is the moment photography stops being only a recording medium and starts being — finally — an art with a clear job that nothing else can do.
The same way Monet, Van Gogh, Picasso, Kandinsky, and Pollock were not the end of painting. They were the moment painting stopped pretending to be a camera and became something only a human could make.
Use AI as a sketchbook. Keep the camera for the moment.
ZSky AI is built by a working photographer. Use it the way Degas used a camera and Hockney uses a scanner — as a tool for previsualization, moodboards, and ideas you would not otherwise reach. Free to start, no credit card required, unlimited image and video generation (ad-supported on the free tier).
Open ZSky AI Free →Related Reading
Frequently Asked Questions
Did the camera really cause abstract art?
It is the cleanest single cause. Before 1839, painting was the only way to record a face, a battle, or a landscape. Once a daguerreotype could capture a likeness in minutes, painters lost their monopoly on realism and had to find a new reason to paint. Within forty years that search produced Impressionism. Within seventy years it produced full abstraction.
Why did Monet stop trying to paint things accurately?
Because the camera could already do that. Monet shifted to painting light and atmosphere — the part of a scene a black-and-white plate could not capture. His haystacks at different times of day, his Rouen Cathedral series, his late waterlilies — these are studies of perception itself, not of the objects in front of him.
What did Van Gogh do that a photograph could not?
Van Gogh painted emotional truth rather than visual truth. The yellow of a wheatfield in his work is not the yellow of the wheat — it is how the wheat felt to him. A camera records what is in front of it. Van Gogh recorded what was inside him while he stood there.
How did Picasso break realism?
Picasso and Braque invented Cubism by abandoning the single fixed viewpoint. A photograph shows you a face from one angle. A cubist portrait shows you the face, the profile, and the back of the head at the same time — closer to how memory and perception actually work.
What was Kandinsky's breakthrough?
Kandinsky removed the subject entirely. His 1910–1913 work is the first full abstract painting in the Western tradition — color and form with no object behind them. He treated paintings the way composers treat music: pure feeling, no narrative required.
How is Pollock related to all of this?
By the 1940s, photography had become the universal record. Pollock answered with the most un-photographable painting possible: the act of painting itself. His drip canvases are records of his body moving through space — a process a camera cannot reproduce.
Is AI doing to photography what the camera did to painting?
Yes, on the same time scale and for the same reason. AI now produces flawless technical realism faster and cheaper than a working photographer can. Photographers are being pushed toward what only a present human body in a real moment can do — witness, presence, and the unrepeatable second.
Should photographers fight AI or use it?
The painters who survived the camera did both. Degas, Sickert, Bacon, and Hockney all worked from photographs — they used the new tool while pursuing what only painting could do. The photographers who matter most over the next decade are likely doing the same thing now: using AI as a sketchbook, while still walking out the door to capture what AI cannot — the moment that actually happened.