Propaganda, Censorship & Art in the Age of Artificial Imitation

I know, I know, another think-piece on AI. Since the middle of 2023, AI has been somewhat inescapable, and while some already claim, that the AI bubble might be about to burst, that doesn't necessarily mean, that it's going away. I don't want to speak too much on the technical aspect of the technology, though that is certainly interesting, there are more experienced and eloquent people who have already tackled it with more time and better presentation. Instead, I would like to follow up on a sentence whose precise origin currently escapes me: "It's somewhat disquieting that we have decided to first automate things that are expressions of the human condition"

Art has always been interesting to define and think about, because all cultures that want to actively engage with it, will likely want to exert some amount of control over art, its distribution, and subsequently its production. This has been true in the age of the "great classics", wherein those men (and at the time, it was mostly men) whose names we remember today, where commissioned by a patron, whose money is more often than not founded in some variant of nobility. This doesn't make the art produced any less great, though of course one might wonder, what any one of those artists might have produced, had they had complete control over their own output.

Going back even further to ancient cultures, texts and paintings, there were (more or less) strict religious guidelines on how gods and certain events were meant to be depicted. This isn't usually questioned as much by modern audiences than similarly strict guidelines on whether to depict a god at all, as it would become the norm for the Abrahamic religions (with the Christians arguably getting around the issue via the holy trinity).

A lot of things have been said about the role of art in society, either as a reflection of the cultural and political Zeitgeist, or a trojan horse to smuggle philosophy inside of, and while it's certainly both, the actual "intended function" of art seems of course to differ depending on the society it's meant (or permitted) to exist in.

The Worker & Kolkhoz Woman is perhaps the most famous example of a Socialist Realist Art Installation. Picture taken by Andy House Photos

Socialist art has a somewhat notorious aesthetic, that leads itself to be mocked for the alleged "groupthink" that is supposedly a systematic requirement for a collectivist political system. While that is certainly an uninformed reading of the bodies of work of the artists in question, whose names in turn tends to get summarily ignored, unless they defect to capitalist countries, it's not entirely unfounded, and happens to illustrate my point. Socialist realism is probably most known for its representative works out of the Soviet Union, post-revolutionary China, and current North Korea (usually in developing/purposely under-developed nations).

It stands somewhat in contrast to other forms of "realism" in the art world. Whereas realist literature (e.g. The Adventures of Huckleberry Finn) aims to depict an objective truth of events and separate itself from speculation and fantastic/divine interference, socialist realism is meant to depict a socialist ideal, which is supposedly possible, or even within reach, provided everybody (or at least the vast majority) does their part. For a not-so quick primer, check out this episode of Proles of the Round Table, and the literature listed in their sources. Ways of Seeing, by John Berger specifically is a comprehensive source on art throughout history. This in itself is the product of a long struggle with the role of art in Soviet and Soviet-Inspired countries. The context that one should keep in mind for this, is the function art had in the European feudal aristocracies. Art during these times was usually in service of worship, be it within religious context, or direct worship of the aristocracies. As such, they usually depicted religious scenes, people, or notable events, which in themselves would likely be embellished slightly in favour of whichever character one might have been supposed to feel sympathy towards. This would have made the function not quite dissimilar from the initial position that Socialist Realism would have occupied, with the difference that the latter was meant to be aspirational, while the former might have been so only by implication. In this context, art might (and in my opinion should) be read as a form of propaganda. A portrait of a king is meant to project wealth and classical nobility (as in possessing power by birth-right), some of which might be supposed to be projected onto their subjects - usually by the use of symbolic colours, family crests, and other symbols (think the royal lily, or all of the lions) - while the "Worker and Kolkhoz Woman" statue is meant to display a nature of a people more directly in the symbolism of the hammer and sickle. Socialist Realism then stands as a contrast to the traditional art, using the tools of the traditionalists, which were perhaps easier to market to the masses, than other contemporaries, such as the Russian Avant-Garde. This was an explicit consideration by leaders within the Communist Party several times, which is of course repressive, but also not entirely incorrect in consideration. Further history on this is a topic for another time, since it's nuanced and complicated, but suffice it to say that the use of art as propaganda is perhaps as old as the modern conception of art itself.

By the time the Soviet Union wasn't actively enmeshed in wars on its own territory, the second World War had ended, and most of the Soviet Union's neighbours, as well as their allies abroad, had moved away from the quasi-medieval idea of art. Abstract Art (we will return to these guys later), had developed from the Impressionist stepping stone, and arguably supplanted the Realists in popular art culture. Socialist Realism, too, had changed. Taking Soviet Cinema for example, had found its own library of techniques, which are now as present in modern cinema as punctuation is in language (see: Khuleshov, Eisenstein). While the decision to restrict art to the Socialist Realist style by the Communist Party did also hit the film-makers, both the traditionalists of the party, and the artists moved art to other functions, all the way to very direct criticism of Soviet Society (See: Moscow Does not Believe in Tears). Since these changes all happened in the span of post-revolutionary Russia, basically after the 1930s attempt of "resetting" the art produced for the proletariat to a level that would be directly legible to the broad citizenship, the influences of the Socialist Realist style, was obviously very present, even without the explicit need to enforce it, as can be seen by the likes of Tarkowski, who defected from the Soviet Union, but whose style and themes still shows the influence.

Now that we've got the primer for Socialist Art as propaganda straight, and we're hopefully all accustomed to looking at art through one eye, and propaganda through the other, let's move Westward. That the art produced by Riefenstahl and her contemporaries in Germany and abroad was a form of propaganda is hopefully not up for debate, and arguing why this movement was capitalist in nature is a topic for another time. Instead, I'd like to skip past that chapter in Europe's history and look at the post-war art movements, formed as a direct consequence. Most of these art movements seem like echoes of Dadaism and other anti-art movements, itself often a direct response to the horrors of WWI. The most known ones fall into the direct vicinity of abstract expressionism, minimalism, and colour-field paintings. Many of these had also been formed in direct reaction to WWII in one way after another. Rothko for example, approached many of his later pieces with the intent to express different facets of grief and tragedy. Abstract Expressionism dispensed with traditional themes with which one could possibly mythologize anybody specific (or at least that was the idea). This could be read as an attempt to fully divorce art from propaganda. If this was the explicit intention, then it did not work.

American Pop Art was developed from the individualistic lean of abstract expressionism. Those tendencies already elevated the individual artist to celebrity status, most prominent perhaps Pollock, and later Warhol, and later it would put the focus mostly on the individual. Much of pop art depicted or included products, or facsimiles of products that could be bought on the consumer market. While this might have put the meaning of the piece on the individual, it did introduce a framing to the piece that could not be divorced from the piece itself. Perhaps the worries of the CPSU were not entirely unwarranted in that effect. The export of American Pop Art would put the focus not only on the objects depicted, but also the lives that these objects existed in. There is a sort of exceptionalism in the view that such a sphere introduces to the viewer, though that was almost certainly not the intention. By this point, art had become comfortable with choosing its subjects from among the common experiences, and still finding something interesting or intriguing in them. While a full frontal picture of canned soup might not strike the observer as special, if they are already accustomed to the item, know its societal implications, and perhaps even a few details about its manufacturing, to those without this context, it might as well have been advertising, not for the product itself, but rather for the American way of life as it was projected internationally. That way, the picture of canned soup can become part of the perception of the "American Dream".

Meanwhile, US cinema had introduced its own guidelines for newer forms of art. Movies, as by that point perhaps the most immersive medium became subject to the regulations that almost every form of entertainment had to undergo when it was new. The Hays code was introduced in 1937 by what later became the MPAA to defend the studios from the attacks by puritanical organisations and conservative political forces, who introduced local censorship laws, which made it difficult to ascertain which moves could be shown by the time they began production. This would involve a number of censorship guidelines, addressing the content of a film, along with its own mechanism for institutions to make sure they were enforced - initially marketed as an opportunity to dispense with the as of then somewhat unpredictable film censorship boards. In truth, cinema - as one of the primary mediums for cultural export (what would be later ratified as the "Hollywood Industry") - was never not meant to be propaganda. While the Hays Code was eventually overcome, it was first subverted, similarly to the way that the supremacy of Socialist Realism was subverted in the Soviet Union during the same time.

Subtext became hugely important in reading art during this time, seeing as the artistic communities were usually (and stereotypically) more progressive than the system they produced art for. Movies produced under the Hayes' code would be frequently analysed for subtext explicitly, with the idea that one would find it more difficult to censor things that one could argue wasn't there at all. For all intents and purposes, this worked, perhaps better than one would have expected. This reading of artistic works for subtext was often retroactively applied to many works that had established themselves into the (often literary) canon, so a variety of effects. While I believe that this is technically a healthy exercise in media literacy, viewing works primarily through subtext opens up the analysis to an over-reliance of aesthetics.

In a brief humorous interlude, I would like to point out the retroactive homo-erotic subtext that Soviet Realism gained, because much of their symbolism was very heavily gendered. The 'Worker' was often assumed to be male, so a depicted unity of workers often involved muscular men with shapely faces standing close to one another, smiling into the middle distance. Considering Stalin's line on gender theory, it's pretty safe to say this wasn't intended, but I'd argue that perhaps this could be embraced by future purveyors of Socialist Realism.

Aesthetics and art have a difficult relationship. Depending on the day, I might be tempted to describe it as the "packaging" that the themes come in. It's undeniably a very important part of art, though interacting with, and through aesthetics can lead to some cargo-cultish results. This effect can be enforced, when the artwork is somehow divorced of its original context, such as it often was with pop art. This isn't unique to pop-art, nor was pop art a form that was more suited to this removal of its original context. One might argue, that it only really worked exactly in the context that it was not only presented, but also created in. However, it did flourish in a time, where mechanical reproduction of artworks for use in personal decorating was on the rise. This can be problematic in several aspects. Most obviously, this can lead easily into a variety of "bad tastes". Notoriously, members of the nouveau riche are regularly accused of "cosplaying wealth", through the overuse of status symbols such as gold colouring or over-reliance of brands, and the infamous McMansions. It becomes more complicated in cases where the subject matter is overtly heavier than just a question of capital. Such things can easily be read as culturally insensitive, or even disrespectul of historical atrocities. For example, one can still acquire replicas of SS uniforms as Halloween costumes, though that does seem like an overt invitation to be bullied. It works the other way around as well: Using aesthetics to mask content that is actively harmful, or that is otherwise assumed to be forbidden. Before Richard Spencer got punched in the face on live television, many journalistic reports on him couldn't help themselves from pointing out how well he dressed, and how surprisingly good he looked. His aesthetics gave him a platform, a tactic still employed by many of the modern fascists in politics today, and isn't remotely picked up on as often as would be necessary to claim any sort of political proficiency on account of voters, statespeople and journalists alike.

The spectre of Walter Benjamin's "Art in the Age of Mechanical Reproduction" looms large, now that reproduction has moved from mechanical to largely digital. Digitizing the tools for the reproduction of artistic works has democratized first the reproduction, and also the creation of art as a whole, which isn't technically a negative. It has somewhat trivialized the process of reproduction, at the very least, to the point where there are now entire industries dedicated to the prevention of digital reproduction. Think of the concept of Intellectual property on the Internet, and the amount of otherwise impressively unsuccessful lawyers making a living off it. Further, the entire concept of NFTs follows the need to reverse this trivialization of digital reproduction, and you will find enough people ticked off by the concept of screenshotting an NFT to at least confirm the instinctual rejection of the idea that the genie can't be put back in the box. It might certainly merit a closer look another time, but given the size of the above introduction, I suspect that desperately needs to be tabled.

Instead, let's return to the introduction paragraph, and remind ourselves of every description of Machine Learning we've read: A machine learning model takes a dataset, and guesses how to categorize. If the result is correct, it's rewarded, if it's not, it's punished, so that next time around, its guess will be closer. Do this enough, and the model will eventually produce the desired result. Beyond the question of intent, something that has been used for qualifying, and setting apart "human-generated" from machine-generated art, something else should occur to the reader. Independently of the intent a "prompt-engineer" however legitimate one might believe that profession to be, where does it differ from aesthetics and do they fit? As already mentioned, there need not necessarily be a difference between the two, and mixing them for the purpose of propaganda is not only easy, but almost unavoidable. Like in the previously discussed examples, it matters then primarily who uses it and for what purpose.

I chose pop art previously not because of a special like or dislike of the movement. I don't have a thorough enough understanding of its nuances to feel strongly either way, beyond the aforementioned respect for certain individual artists. I chose it because what happened to it, when it left the US. The style of American Pop Art would worm its way into global advertising, and the style became short-hand for the consumerist behaviour that is now mostly subject to criticism. Advertising itself, which had been difficult to divorce from the aesthetics of early pop-art, due to the matter-of-fact simplicity it had in its early days, would quickly abandon that style of selling its subjects, while still retaining the several approximate aesthetic derivatives of commercialized realism that pop art pioneered in the art world. In that scope, it's been observed several times over how "mere" imitation can dilute not only the point of the imitated art-work, but also the viewer reaction. Otherwise, one might argue that advertising wouldn't have had to take the absurdist turn that preceded its current politicized lean. Tasking AI with the generation of art is almost certainly a fast track to the worst and/or blandest examples of the imitations each "era" of advertising has produced, because in a way, their processes are very similar. While advertising still currently deploys artists, they are often somewhat detached from the product they are tasked to market. One can perhaps most often notice this in music in advertising. If it's not a cacophonous soundscape already, the choice of song played in the background might be recognizable, but without a clear connection to what the product is selling. For example, Pachelbel's Canon in D has been used to advertise cars, hair products, and vacations, none of which are remotely topical. The selection might as well be random. Taking advertising as an orientation for an industry of AI generated art, then projecting the industrialized production of advertising onto the social place that art occupies paints a somewhat bleak picture.

Advertising itself is already a form of propaganda anyway, and it was so even before the products purported to contain political and social values. Think of the stereotyping of "German Automobiles", "French cuisine" or "American Firearms". Each of these industries wants to market itself to require some qualities, made inherent by attaching a Nationality to it. Not only is this pervasive, and like most propaganda, at best partially true, this propagandistic slant is the point of each individual advertisement, as opposed to the structural byproduct of a movement/industry. When imitation through AI then becomes the chief producer of art, the similarities between the resulting art-works and advertising, along with the removal of the link between the aesthetics and the artistic content is a recipe for disaster, not to mention the labour implications.

There's good news in here somewhere. The way these models are trained still require human input. I, for one, don't really advocate for replacing every job - especially every artist's job - with "prompt engineers". I feel like that would make the human existence mind-numbing in the most boring way, by reducing the collective human experience into the experiments with a million monkeys on a million typewriters, with the depressing difference that the end product isn't Shakespeare, but "cartoon, picture of cartoon monkey, cartoon mountain in background, bright colours, bright green, bright colors, featured on deviantart". These tools can expedite the processes that nobody people aren't really interested in doing for themselves. This definitely does not work without other changes. In a way, those people arguing that AI can be a tool for the creation of art like any other are correct - just not in the way they propose it being rolled out. This step should not be new, and while industry and capital have never really cared for the workers and replaced them relatively quickly after the technology could be easily reproduced, perhaps doing so in perpetuity is not something one should be proud or indifferent of. There's several ways to participate in exploitation, and silent acceptance is one of them. AI is interesting technology, I think most people wouldn't really argue against that. However, its development is not essential, and some toys only get to be used, when the owner can demonstrate they know how to use them. It's really no different from a driving test, albeit with slightly more collective passing requirements. A first proposal might look like this:

- A company must employ a certain number of workers every financial year, the percentage of industry rookies should reflect the number of graduate or pre-graduates recorded in that year.

- New industry hires are not, and should not be expected to turn a profit, until they have learned the craft, and can do work that can't be automated

- All employees are entitled to a living wage (obviously) and no difference in payment should be made in compensation between those still learning and those already accustomed to the work

- AI contributions must be vetted extensively. The vetting process does not qualify as human contribution to the art-work, in the same way the review of a critic does not qualify as their contribution to the art-work

- The contribution of AI must be made clear by the companies generating, commissioning, and distributing both the model, and the work.

- The price of fully AI generated art should be tied a set percentage above the electricity cost after training the model (proposing 2% as a start, but I haven't given it much more thought beyond not feeling comfortable with more)