Syllabus for the Internet Telematic Society

On Vilém Flusser‘s philosophy of what a camera wants, and what such a perspective misses

SYLLABUS FOR THE INTERNET is a series about single books or bodies of work written prior to the rise of the consumer internet that now provide a way to understand the web as we know it today. View the others here.

In 1983, the philosopher and essayist Vilém Flusser wrote a book called Towards a Philosophy of Photography. In less than a hundred pages, he sketched out a bizarre and counterintuitive definition of photography as a kind of world-spanning artificial intelligence, using human beings as tools to realize its sole directive: the production of more, and more varied, photographic images. In later works, Flusser expanded these ideas into an account of what he called “telematic society”: a new social order organized around the circulation and consumption of digital images across a global network. For him, this was a speculative vision of the near future; for us, it sounds uncannily close to the world we know today. 

For Flusser, philosophy was a form of science fiction. He studied what he saw as the fundamental tendencies of his contemporary society to predict the sort of world his readers would soon be living in. While his interests were broad, the central proposition of his work was that the characteristics of human societies are strongly influenced by the media technologies available to them. Photography, he believed, was the herald of a revolutionary transformation in the organization of global society, soon to be fulfilled by the creation of what we would now call “the internet,” although Flusser does not use the term himself.

Flusser saw photography, in its essence, as a global information-processing system

The kinds of things that Flusser had to say about photography have little in common with the canonical theory that preceded it, like Susan Sontag’s On Photography (1977) or Roland Barthes’s Camera Lucida (1980). Flusser had no interest in analyzing the style or content of any particular photograph or photographic oeuvre, and he didn’t much care about the ethics of taking or consuming snapshots. Neither was he bothered with describing the specific role photography played in the news industry, or advertising, or fashion. He saw photography, in its essence, as a global information-processing system, and as far as he was concerned, its most important effects were best measured on a global, systemic scale. 

Flusser writes from a standpoint of extreme abstraction, a position that can be explained, to some degree, by his biography. Born in Prague in 1920 to a family of Czech Jewish intellectuals, Flusser had just begun university when the Nazis invaded in 1939. He elected to flee; most of his family remained and were killed in the Nazi death camps. A year later Flusser emigrated to Brazil, where he found work writing newspaper columns and teaching at the University of São Paolo. Over the next 20 years, he committed fully to the cultural life of his adopted country, but in 1964, the Brazilian government was overthrown by a military coup, and the intellectual communities in which he participated were stifled by repression and censorship. He returned to Europe in 1972, settling in the south of France where he produced the texts for which he is chiefly remembered: the media-theoretical trilogy of Towards a Philosophy of Photography (1983), Into the Universe of Technical Images (1985), and Does Writing Have a Future? (1987). 

Flusser’s life, as he described it, was defined by the experience of “groundlessness.” The destruction of his home and the murder of his family left him, as he saw it, without a nation, a culture, or a history, stripped of the implicit frameworks that people who are “grounded” use to make sense of their existence. But even as he mourned these losses, he also saw in them the possibility of a new, radically cosmopolitan philosophical standpoint. “Regardless of being from Prague or London,” he wrote, “one is provincial if one has grounding. But whoever has been removed from the order of things can see the whole world.”

Flusser’s work is defined by his attempt to locate this “view from above,” to discover fundamental principles that can illuminate all other social phenomena. More than anywhere else, he found this in media. The function of modern media technologies, as he saw them, was precisely to provide an overarching societal infrastructure for communication. Other social forces — cultural, economic, demographic — may have their own autonomous logics, he allowed, but to the degree that they must pass through this media infrastructure they are profoundly shaped by it. Hence, his writing focuses almost exclusively on this infrastructural layer and has almost nothing to say about the traditional concerns of photographic criticism: gender, colonialism, class, criminality and so on.

Flusser’s life was defined by “groundlessness”; Flusser’s work attempts to locate this “view from above”

As you might imagine, this approach produces a theory with some seriously problematic blind spots. Having said that, it must be granted that from this elevated vantage point, Flusser does seem to have seen a little further than most. His ideas prefigured the internet — mobile computing, social media, platform economies — with eerie precision, and he repeatedly anticipated our current debates around issues like work, data, and post-capitalism. The logic by which he arrived at his predictions was frequently wrongheaded, and the conclusions he drew from them often flatly incorrect, but this combination of perspicacity and blindness — and what the tension between them now reveals — makes him uniquely relevant today. Read with a skeptical eye, Flusser’s work still has a lot to teach us about the power and pitfalls of techno-futurism.

When a person takes a photograph, they make a series of decisions as to the scene they wish to depict and how they want it to appear. These are free choices, but they are made within categories defined by the technical properties of the camera: distance, angle, shutter speed, focal length, color, and so on. They are also influenced by cultural scripts that dictate the proper subjects for photography — a smiling face, a picturesque landscape — and how they should be shot. Altogether, these technical parameters and cultural scripts form what Flusser called “the photographic program”: a comprehensive set of rules for the production of photographs. 

Photography, as Flusser saw it, is a game we play with and against the camera program. Each photograph we take is a realization of one possible combination of the program’s rules; the pleasure of the game is in exploring and experimenting with this near-infinite range of “photographic possibilities.” Photographers can try to “exhaust” the program by realizing as many different permutations as they can (an impossible goal, given the near infinite possibilities implicit in the camera program) or they can attempt to “outwit” the program by discovering an original combination of parameters not already contained in it. This could mean finding an original subject for photography (images of the storms of Jupiter produced by orbiting spacecraft, for instance), or a new way to approach a familiar subject (think of how the selfie has changed the way we shoot faces). But this provides only a temporary victory, since any novelty produced is immediately absorbed by the program. 

What Flusser describes is an AI whose prime directive is to produce as many different photographs as possible

For Flusser, the rhythm of the photography game is not defined by progress toward a final goal but by a recurring sequence of button presses, each releasing a discrete dopamine hit, and a constant stream of redundant images created in pursuit of unrealized novelty. If this sounds familiar, that’s because it also describes the engagement loop of contemporary social media.

While individual photographers find themselves trapped on this treadmill, the photographic program steadily improves itself, using their activities as feedback. The more people produce photographs, the more combinations of rules the program comes to contain. The more complex the program, the more compelling the game, and the more people are encouraged to produce photographs. Photography, Flusser therefore concludes, is a self-optimizing information-processing system. Or, to put it another way, an artificial intelligence.

What Flusser was trying to get at here is similar to the infamous “paperclip maximization” thought experiment proposed by philosopher Nick Bostrom (and later immortalized as a browser game by Frank Lantz). In Bostrom’s fable, a super-powered artificial intelligence is programmed to produce the maximum amount of paperclips as efficiently as possible and ends up turning the entire solar system into a vast paperclip production line, with no regard for human well-being. In Flusser’s account of photography, it similarly becomes “a means of programming society — with absolute necessity but in each individual case by chance (i.e. automatically) — to act as a magic feedback mechanism for the benefit of a combination, and of the automatic reprogramming of society into dice, into pieces in the game, into functionaries.” 

That is, what Flusser is describing here is an AI whose prime directive is to produce as many different photographs as possible, employing whatever means it can to execute this function for the lowest cost. To this end, it recruits human beings to serve as its fleshy prostheses, employed to point the camera, press the shutter, and create photographs in such volume that a certain percentage will inevitably, by accident or by design, contain “new information” to be absorbed into the program. The various motivations that people might have for making photographs — to amuse, to titillate, to prove, to shame, and so on — are meaningful to the program in only one respect: How successful are they in stimulating the creation of novel photographs? Otherwise, the camera program is, just like the paperclip maximizer, completely indifferent to human interests and values.

It’s worth taking a moment to appreciate the startling degree to which this describes how photography functions in contemporary capitalism: Flusser’s abstractions have been realized almost point for point in the internal plumbing of Facebook, Instagram, Google, and the multitude of machine-learning algorithms that feed off constant streams of image data. Most photographs produced today are, from their inception, part of a network in which they are circulated and their reception monitored for feedback to drive user engagement. And while people make and enjoy photos for any imaginable personal reason, these are as irrelevant to their value for their aggregators as they are to Flusser’s “photographic program.”

Today, the most tangible manifestations of the “photographic program” are at the surreal fringes of AI-enabled visual technology. Think, for example, of Clearview AI, the facial-recognition company that scrapes images from social media to train algorithms it markets to law enforcement. If for any reason a picture of you has ever been uploaded to the internet, it’s probably in the Clearview database, which means that not only are you legible to its search engine but you’ve also been used to refine it. 

Meanwhile, software like Dall-E — the impressively resourceful text-to-image generator from OpenAI — seems to demonstrate that the photographic program has absorbed such a quantity of images that it no longer needs photographers to input new material. A Pictionary-style prompt is all that’s required to produce an original combination of visual elements. The sharp, pristine images that have so far been published on behalf of Dall-E 2 prefigure a world where the production of images has become fully automated. Yet the smeared textures of the images produced by Dall-E Mini (the publicly available version of the tool) help remind us what they actually are: the synthetic pulp of millions of pictures that were once uploaded to the internet for some other purpose, chewed up, and regurgitated by a neural net. 

Neither Clearview AI nor Dall-E, much less Facebook or Twitter, would have been a surprise to Flusser. Looking at the personal computing technology of the early 1980s, he imaginatively extended it into a global circuitry, linking “person to person, a bit like nervous pathways and nerve cells join together” to form a “global brain.” In other words, he anticipated something like the internet, but not so much as radical new technology than as a logical evolution already implicit in photography. The photographic program would, he believed, be both the blueprint for the next generation of media technologies and the structural model on which the society of the future would be based.

Flusser’s term for this new order was “telematic society,” an amalgam of “telecommunications” (the technologies which would be crucial to social organization) and “informatics” (highlighting the crucial role that information would play in its political economy). Telematic society would be defined, he predicted, by a condition of simultaneous isolation and interconnectivity: “people will sit in separate cells, playing with their fingertips on keyboards, staring at tiny screens, receiving, changing, and sending images … [forming] a dialogic net, a global superbrain.” 

Telematic society would be defined by simultaneous isolation and interconnectivity; life in such a society would be equally boring and chaotic

The function of this “superbrain,” just like that of the photographic program, would be the constant creation of novel images from a set of pre-existing parameters. History, politics, current events — all would be grist for the production of new visual spectacles, and every individual would be a player in an unlimited game, competing to find the most eye-catching way to repurpose and circulate old cultural matter. The organization of this system would be governed by the “feedback loop between image and receiver,” which enables “the images to change, to become … more and more like the receivers what them to be,” while in turn the receivers “become more and more like the images want them to be.” As in Flusser’s model of photography, this system would absorb the idiosyncratic desires of individuals — for knowledge, amusement, connection, status — into its programmatic engagement cycle of input, reward, and feedback.

Life in such a society, Flusser imagined, would be boring and chaotic by equal measure: boring because the feedback loop between image and receiver would reduce culture to a state of bland homogeneity and redundant replications of old content; chaotic because a society dedicated to the constant production of novelty will move so quickly that all sense of linearity would be lost. “Events,” he wrote, “are caught in the undertow of images and roll against them more and more wildly. One political event follows another more and more precipitously, a scientific theory is introduced, an artistic style replaces another almost before it has been established.”

No doubt many of us can see the contours of our current situation — the often isolating and anhedonic nature of online life, the increasingly cannibalistic nature of pop culture, the general sense of accelerating crisis without resolution — in this schematic outline. At the same time, there are many aspects of modern life that Flusser got almost exactly wrong. In many ways, these errors are more important than his accurate predictions. Not only do Flusser’s gaffes highlight the blind spots in his theory, they can help us see the risks inherent in his style of seductive techno-prophecy.

The problems with Flusser’s ideas really come to the fore with his account of how power would be exercised in a telematic society. He saw two possible future paths: “One moves toward a centrally programmed, totalitarian society of image receivers and image administrators, the other toward a dialogic, telematic society of image producers and image collectors.” 

In the “centrally programmed” version, images are generated in some sort of state or corporate headquarters before being broadcast throughout the system. Feedback is likewise centrally collected, and whatever useful insights or tools might be derived from it remain in the hands of the elite. In the “dialogic” version, images are shared freely across a horizontal network, and all nodes participate equally in their creation and transformation (a vision not dissimilar, incidentally, to some of the utopian imaginaries currently emerging from the blockchain trend).

Flusser believed contemporary culture was predominantly “centrally programmed” and that if this were not countered by a strong “dialogic” tendency, then the telematic society would be fundamentally “fascistic,” dominated by centralized “senders” with absolute control of the content flowing through the network. To avert this dystopia, he argued, a network in which each individual participant can create and circulate their own images and receive and implement their own feedback would be necessary. This structure would allow for political affinities to emerge spontaneously from the aggregate of each individual’s decisions, ultimately producing what he called “a broad, worldwide consensus relating to the programming of apparatuses” — the prerequisite, he believed, for the democratic administration of the network.

What Flusser seems not to have anticipated is the situation we have actually ended up with: a worst-of-both-worlds combination of his two models. The creation and circulation of images has indeed been delegated to individual nodes, who receive feedback as likes, shares, subscriptions, and so on. But we also have centralized “image administrators” in the form of platforms like Facebook, Google, and Twitter, which have effectively monopolized the ability to take a global overview of the “program.” The ability to create and broadcast images is not, it turns out, as significant as being able to watch and learn from how they flow through the network. 

What Flusser seems not to have anticipated is a worst-of-both-worlds combination of his two models

Flusser’s inability to foresee this eventuality stems from his attitude to technology and power. He was convinced that the revolutionary possibilities of media would obviate older models of political analysis. In a telematic society, he imagined, the majority of manual labor would be automated, meaning that most people would be employed in the service industries or what we now call “knowledge work.” In this condition, he argued, the old Marxist categories would be out the window. People would no longer be “workers” but “functionaries”; power would be held not by the owner of the means of production but “the programmer of the information”; the social order would no longer determined primarily by class interests but by the inherent programmatic logic of the network itself, over and above its socioeconomic context. He criticized the work of the Frankfurt School for trying to locate the “secret, superhuman power” of capitalism behind the curtain of modern society, “instead of taking it for granted that the programming proceeds in a mindless automatic fashion.” 

To some extent, this was trollish provocation on Flusser’s part. But it does reflect his genuine belief that the programmatic dynamic he’d discovered in photography — a self-directed force that operates on its own, independent of the competing interests which facilitated and sustained its development — would be decisive in shaping the society of the future. History has not exactly proved him wrong, but if the experience of the past 40 years has demonstrated anything, it’s that technologies in themselves do not drive revolutionary change. At most, they remediate or mutate pre-existing social and historical dynamics. 

The really salient problem — which Flusser is ultimately unable to address — is how political economies and technological formations are mutually constituted. Under what conditions do new technologies transform or destroy old social forms, and vice versa? Read against the grain, Flusser can actually help us address these questions. In his work, qualities we associate with modern technologies — the development of a global network that connects users via feedback loops, the use of images as inputs for an artificial intelligence — are presented as properties of photography rather than mobile computing or social media. This suggests an alternative framework for understanding contemporary tech, one which sees “the internet” or “artificial intelligence” as part of a much broader history, encompassing not just photography but also finance, workplace management, advertising, and range of other spheres in which programmatic, self-optimizing processes have become increasingly dominant over the years.

Seen from this perspective, the automative social logics that Flusser associates with photographic technology can instead be traced to their origins within capitalism. Marx himself identified an “inhuman power” at the heart of capital, a kind of self-optimizing feedback loop that drives the system toward more intense and profitable forms of exploitation regardless of the interests of any specific individual or group. Building on this insight, other scholars have suggested that there may be an “elective affinity” between artificial intelligence and capitalism. By loosening its dependence on human labor and decision-making, AI accelerates the automative tendency already implicit in contemporary capitalism. 

Beginning with this kind of framework, the aim would then be to explore how technologies like photography have evolved in concert with an increasingly automated and programmatic capitalism and the role they have played in preserving its core logics of domination, even through periods of deep political and social transformation. 

For all his flaws, Flusser is still worth reading today. His writing is brisk, pithy, and expository, displaying an eagerness to make himself understood that his more famous contemporaries in media theory did not always share. Composed as sequences of short essays, his texts are intended to pique and provoke, and this is how they are best engaged. His bold, and frequently accurate, predictions are a benchmark to measure our ideas against; he challenges us to do a better job of accounting for the trends he foresaw, as well as those he didn’t. 

At the same time, his work crystallizes the dangers of a mode of techno-futurism based on revolutionary leaps, abrupt paradigm shifts, and scorched-earth transformations. After all, one of the distinctive characteristics of AI-enabled image archives is the way they shackle us to the past, both by embalming old aesthetics for future reuse and by recapitulating pre-existing structures of racial and gender injustice through, for instance, facial recognition algorithms. If yesterday keeps repeating on us, this is partly a consequence of the amnesiac thinking that Flusser’s work ultimately advocates. He’s the opposite of Benjamin’s “angel of history”: Eyes fixed unswervingly on the future, Flusser loses sight of the storm that’s still propelling us forward. 

Richard Woodall is a writer and researcher based in Sheffield, England. His interests include the history and politics of visual technologies, as well as the analysis and critique of digital capitalism.