In late 1918, the newly created Weimar Republic emerged at a moment in time that would coincide with drastic changes to Germany’s culture as a whole, with a particular focus on its visual culture. Photography had already been invented much earlier (at a time when Germany was still a patchwork of statelets). But it had taken time for what we could now consider its most defining feature to fully emerge: mass reproduction.

Mass reproduction relied on the technology being available. The moment it became possible to produce photographs using ink on paper, they would be used by newspaper and magazines. While the process had been available before the advent of the Weimar Republic, the disappearance of the imperial regime and its staid, backwards conventions opened up a way for innovations of all kinds to flourish.

In August 1919, Berliner Illustrierte Zeitung (a news magazine) published a photograph of President Friedrich Ebert and Defense Minister Gustav Noske on its cover. Ebert, a social democrat (when that term still meant anything), had become President just three days earlier, taking his oath to a new constitution that had been created not in Berlin, the country’s capital that was still beset by strife, but in sleepy Weimar (past home of national hero Goethe and future home of national disgrace Buchenwald), giving the new Republic its name.

But the photo the editors had chosen was not one that could have been published under the previous imperial regime. Ebert and Noske had been photographed at the beach. They’re seen knee deep in water, posing somewhat awkwardly for the photographer (there is a third man in the picture who is mostly submerged and who holds up a pitchfork — clearly intending to play the role of Neptune).

Ebert and Noske’s paunchy figures and swimming trunks do not necessarily make for the most flattering look — this much is clear even today, over 100 years later. It’s difficult to fully assess the impact the photograph and its publication might have had at the time, though: ours is a world flooded with photographs, and we have seen our leaders in all kinds of flattering and unflattering situations. The Weimar Republic citizens had not.

In fact, Germany’s last emperor, the hapless Wilhelm II, had been born with birth defects, which included not only mild brain damage but also a left arm that was noticeably shorter than the right one. If you know about this, you can see the man’s efforts to conceal the fact in official photographs. But it would have been unthinkable for a news magazine to publish a corresponding photograph such as the one on the cover of Berliner Illustrierte Zeitung in August 1919.

Before 1918, it would have been unthinkable for an artist to take the cover photograph, use it for their own art work, and show it. But that’s exactly what Hannah Höch did when she created Staatshäupter (Heads of State), one of her earliest collages (you can find it included in this article). She cut out the two figures from the Berliner Illustrierte Zeitung cover and put them on top of an embroidery pattern.

At the time, Höch was engaged in the world of dada. But Staatshäupter is not dada: it’s a clearly political piece of art, created in response to the photograph on the cover of a news magazine but probably also in response to ensuing discussions around it.

One could view the publication of the Ebert/Noske photograph on the cover of a news magazine as an explicit criticism of the newly formed republic. After all, the dignity of the office of president was hardly reflected therein. One could also view the photograph as the beginning of a new era, an era of democracy where even the country’s leaders were like the rest of us, enjoying their time at the beach and possibly not looking presentable according to the old conventions that took a little longer than the emperor to leave the country.

But Höch’s work of art was something else. Replicating either one of the two possibilities discussed above would have been too simple. It’s not that doing so would have created a form of agitprop. It would have been the wrong kind of agitprop.

The newly formed Weimar Republic had granted women the right to vote, a first in Germany’s history. True gender equality still remained only a pipe dream for women (today’s Germany has made progress, but it’s still a pipe dream). However, the explicit recognition that women should have the same say in choosing their country’s leaders as men was a clear and long overdue acknowledgement of the fact that women’s roles had been artificially restricted.

By placing the head of state and his defense minister on an embroidery pattern, Höch effectively placed them in her female world. Obviously, much like not all men have powerful roles in government not all women use embroidery. What Höch was playing with with her kitchen knife (to adopt a short phrase from the title of her most widely known piece of art; reproduced in part above) were symbols: photographs and visual materials both as source material for a new object and as stand ins for larger concepts.

In other words and using today’s language, Staatshäupter appropriates source material to create a new, different meaning, a meaning envisioned by the person creating it. And that meaning was tied to a form of criticism only possible in the arts, a criticism that arises solely from the juxtaposition of visuals that ordinarily are not seen together.

You will want to keep these two key aspects in mind — the juxtaposition/synthesis of visual material and the clear intent of criticism.

Given the possibilities provided by photomontage, it is not surprising to understand how this new form of visual culture became widely used in the Weimar Republic. Other artists embraced it — László Moholy-Nagy included it in his treatises around visual arts.

John Heartfield employed montage to devastating effect. Born Helmut Herzfeld (he changed his name as a protest against strong anti-British sentiments during the final imperial years), the artist would end up as number five of the Gestapo’s most-wanted list for his work: a ruthless and biting anti-fascism that had, for example, an anonymous industrialist put large wads of cash into Hitler’s hand.

A lot of Heartfield’s work is agitprop. And yet, since it is critical it is art. For example, his Der alte Wahlspruch im “neuen” Reich: Blut und Eisen (The Slogan of the “new” Reich: Blood and Iron) takes two prominent fascist concepts, blood and iron, to form a swastika out of four butcher’s axes, blood still dripping from their blades. It’s crude, sure, but it’s effective — and if things had gone badly (as they did for some other artists), his art could have cost Heartfield his life (he managed to escape the Nazis).

Many of Heartfield’s montages ended up on covers of left-wing magazines, thus completing the circle that Höch had started. Given the many contradictions and contrasting impulses that existed in the Weimar Republic (and that would ultimately doom it), you could argue that montage was its defining art form. Not photography, not film, no, it was the montaging of images to create new images that exemplified the time — and that would also foreshadow much later, postmodern thinking.

But montage was not only used in the Weimar Republic. Totalitarian regimes such as the Soviet Union or fascist Japan produced publications in which montage played a prominent role. It’s instructive to compare Soviet or Japanese montages with Weimarian ones: the techniques and ideas are the same. But since Soviet or Japanese montages were largely made with the idea of creating propaganda in mind, there is little artistic value to them.

Montage thrives when it is employed in a critical fashion; when it is used for propaganda, it falls flat pretty quickly.

Roughly one hundred years later, montage is not used all that much any longer. But it has not vanished completely. Christopher Spencer (who goes by Cold War Steve) has been treading in John Heartfield’s footsteps for years, using similar techniques while expanding the approach. Often relying on classical paintings as starting points, with today’s tools Spencer could create much more seamless images than he does.

I would argue that the visibility of the montage forms the core of why his images are so potent. The obscenity of much of what is being criticized by these montages in part reveals itself through the crassness with which Spencer splices together his raw material. It’s extremely effective. If I were someone who believed in an afterlife, I would probably imagine Heartfield and Höch smiling about their successor.

That all said, none of the above is the focus of this essay. What I really want to try to understand is something that has been pointed out by Roland Meyer, Professor in Digital Cultures and Arts at University of Zurich and Zurich University of the Arts. Through his writings and social-media presence, Meyer has become my go-to person for discussions around what has become known as generative artificial intelligence (GAI) image making.

A quick note first. I do not believe that the inclusion of the term “intelligence” in that string of words makes sense. Unless you define intelligence in an extremely narrow sense, what these kinds of tools do does not in fact qualify as artificial intelligence. Obviously, there’s a reason why the makers of the tools use the term: it’s to essentially shut down part of possible discussions. After all who wants to argue against intelligence?

Going forward, I will continue using the term for one and only one reason: I do not believe that the arguments I want to make are being served by defining my own term for the tools in question. People such as Roland Meyer are much more well suited to come up with meaningful descriptors; and unfortunately, most of the debates have uncritically used the term.

One of my main grievances around the debates I have seen is that they discuss GAI on its own terms, regardless of whether the economics behind it are concerned (essentially large-scale thievery of other people’s work) or the resulting images themselves. As someone who has a large interest in visual culture, I believe that you can’t discuss GAI images without placing it into the larger visual context. And that context has be larger than merely the contemporary one.

Here then is part of the reason why I discussed montage in so much detail above. I believe that montage and GAI have enough aspects in common for them to be discussed in relation to each other. Obviously, I am not claiming that montage and GAI are the same. There are considerable technical differences; but these differences are largely irrelevant for what I am going to discuss.

In many articles and interviews, Meyer has spoken about the fact that GAI tools serve nostalgia and that they’re amplifiers of cliches. One of the topics that also has come up is the question why GAI tools, in particular image tools, have been so widely used by the far right. Why is that? I have been trying to understand possible reasons for some time.

Many people refuse to use GAI for exactly the same reason that I refuse to use it: it relies on stolen material. But there are two problems with that approach. First, if GAI image making served people outside of the far right just as much, a lot more people would use it.

I do not want to imply that people are callous or don’t care about the problem of the thievery. After all, we live in a day and age where it’s almost impossible to remain pure, given the extent with which neoliberal capitalism has salted the fields. Whether it’s being able to shop for groceries that aren’t wrapped in enormous amounts of plastic or staying clear of bio-engineered food items or only using corporations that do not cause massive harm — it’s very, very hard to steer around whatever one has a problem with.

Thus, at least in theory GAI images could be used to combat fascism. But they’re not — because they can’t do the job. Why exactly is that?

Second, to phrase it in a crass fashion, montage relies on stolen images just as much as GAI. Hannah Höch “stole” images just as much as John Heartfield did. In the past decades, the world of art has spent considerable effort on trying to understand and/or define why or how it can be OK to use someone else’s work. Whether or not I buy all of the ideas and/or explanations brought up so far is besides the point here.

The point is that despite their technical differences, montage and GAI rely on appropriation. I wanted to understand why I’m OK with Hannah Höch, John Heartfield, and Christopher Spencer — but not with GAI images. Isn’t this hypocritical?

I told myself that the answer had to go beyond the fact that one side mostly represents the fascists whereas the other side does not. Even though I have enormous problems with fascism, it felt unsatisfactory to me to remain at that level.

So why does the far right rely on GAI so much?

The first aspect that seems important to me is that contemporary fascism does not have a continuity of visual representation of their ideology. After 1945, the production of such imagery largely ceased because of the huge stigma associated with it. Fascists did not disappear. But in most countries I am aware of there was societal agreement that fascism was not acceptable, whether politically or culturally.

That agreement has now been shattered. Still, you did not have, say, German citizens (whether artists or whoever else) create and celebrate fascist art after 1945. Or rather, there was art that used fascist ideas, but it was not seen and/or discussed in such a context. Essentially, the contemporary far-right has to create their visuals starting from scratch.

In addition, the far right does not have any talented artists to work with. There’s a simple reason: true good art lives from working with uncertainty and doubt, and the far right rejects both of those. The contemporary far right is mostly driven by dissatisfied and extremely insecure males who nurture their largely imaginary grievances among themselves. Those males suffer from being unable to process their own uncertainty and doubt.

For fascists, uncertainty and doubt are signs of weakness — and thus to be rejected. Thus, there can be no true far-right art. Whatever far-right art might exist conforms to very broad conventions of art, but in terms of even the loosest sense of art criticism, it’s laughably bad. If you don’t believe me, just watch this short video (please note: the video is in equal part funny and deeply disturbing).

Far-right ideology expressly rejects proper criticism (whether the art kind or any other) and, instead, replaces it with a phony simulation. Its art reflects this basic fact.

For what it’s worth, you can say the same about any kind of totalitarian ideology. Were the Soviet Union still around, its leaders would still embrace the same restrictions on art. Totalitarian regimes in general do not support freedom of the arts. (If you look at the visual representations employed by Nazi Germany and the Soviet Union, you see very clear similarities: the heroic leader, the very traditional family unit composed of extremely attractive members, the power of the state as expressed through its military might, etc.)

In addition, unlike communism fascism does not have an actual vision of the future. Fascism only offers an imaginary, idealized view of a national past as its ideal. In other words, whatever vision this might be — it has nothing to do with a better future. Instead, the idealized future is merely a rehash of the past.

In other words, fascism does not have to truly imagine what might exist. Instead, it only has to imagine what might exist again, a possibly slightly modernized version of the past (modernized of course only in terms of amenities, not in terms of thinking or culture). Even the imaginary communities envisioned by far-right tech billionaires (whether on Mars, in Greenland, or Gaza) are ultimately little more than the romantic villages from an (again largely) imaginary past.

In an artistic sense, fascism thus does not offer a vision — whereas true art relies on a person’s vision to imagine something different, something better, something that — for the art to be truly good — might not be fully formed and might embody contradictory impulses.

All of the above makes GAI image generators that rely on past, existing images as their source material ideal tools for the fascist project. The many restrictions that most artist rightfully see in that tool are not an actual problem for fascists: it’s not that there’s no need for contradictory impulses, it’s that such contradictions are inacceptable. (It’s absolutely no coincidence that Hitler started out as a really bad artist.)

Thus the nostalgia detected in GAI images by Roland Meyer is absolutely not innocent. By itself, nostalgia is not necessarily a good entity to begin with, given the extent with which one’s imagination shapes and distorts the past one is nostalgic for. But in the hands of the fascists, this nostalgia becomes menacing; and as we’re witnessing all over the world, this nostalgia manifests itself in actual violence.

It is absolutely no surprise that once GAI tools were able to create Studio Ghibli style imagery, American fascists immediately used them to create images depicting violence. Without violence and domination fascism would not exist. And any tool that makes the creation of imagery depicting violence and domination so easy is custom-made for fascism.

On a more technical level, whereas montage not only celebrates but also works with the many artifacts inherent in it — mismatched sizes of people’s figures, say — in GAI images, artifacts undermine them. A GAI image becomes less believable when someone has six fingers on their hand.

In a montage, an artifact is a feature; in GAI, it’s a mistake. Given that GAI strives to avoid mistakes, I can’t see how any self-respecting artist would choose to use it. Good art lives from uncertainty and mistakes — exactly the things the makers of GAI attempt to get rid off.

Lastly, the inherent smoothing employed by GAI that has so many photorealistic images look so similar is mirrored by the largely identical look adopted by the far right. If you watch any Fox News program, say — it is as if the people in front of the camera had agreed to run a smoothing filter over themselves to create a very limited version of what a man or woman might look like. It’s deeply disturbing, of course, in part because it brings Nazi Germany imagery back to life.

As became clear this past week when suddenly the “X” (formerly Twitter) chatbot started going on rants about “white genocide” in South Africa (story; it’s difficult to keep up with this, though, given it has since pivoted to Holocaust denial), many GAI tools have become tools that serve fascism.

It’s clear from what’s going on on “X” (now the equivalent of a Nazi bar) and from the company’s response that the actual problem was not that the chatbot would produce far-right conspiracy theories; the problem was that it would do so in a fashion that made the site and its far-right owner look bad.

It’s doubtful whether Hannah Höch really used a kitchen knife for her work as the title of one of her most important montages implied; what is clear, though, is that she wanted to tell her viewers about the fact a woman had done the work.

Whether it was a kitchen knife or scissors — these tools allowed her full control over what she made. With today’s GAI, anyone using it cedes large parts of such control — and puts it into the hands of people like Elon Musk. A tool might not reveal its makers’ biases quite as openly as the “X” chatbot did; but they’re there, and they are outside of the control of possible users.

In light of so many tech companies having embraced far-right ideas, it’s absolutely clear that GAI tools are not the equivalent of Höch’s kitchen knife. Instead, they’re the equivalent of the butcher’s axes in Heartfield anti-Nazi montage.