Esotericism and Exotericism in AI Art

Using this script, Descript AI will make a video to go alongside it, for free, in around 2 minutes. One can then upload that video to YouTube, for the potential of generating ad revenue and gaining visibility online.

Many inhabitants of the internet with particularly entrepreneurial mindsets have levied this opportunity to produce (rather than create) mass amounts of worthless, generic content riddled with inaccuracies and devoid of the grace of the human mind that has become known as “AI Slop.” These people call themselves “Faceless Youtubers,” which is probably among the more charitable ways one can frame their activities. There are thousands of people making hundreds of thousands of dollars in this way, while making all of YouTube worse for other people.

While this technology is obviously new and therefore too the phenomenon of Faceless Youtubers, they are tapping into something that is fundamentally old and has existed on the internet for quite a long time. Clickbait, low effort content has always flooded the internet for as long as people understood that there was money to be made doing it. Live leak, Popular Mechanics, and YouTube channels like Watch mojo are like this procedure at their base. Like most commercial art, it is exoteric. It is meant for an audience, and we must not mistake the absence of an artist where merely a bad artist resides. While the artist fades into the background for content like this, this is because the driving force of its creation lies in the existence of an audience, not the impulse of the artist to create. Fundamentally, nothing has changed with AI slop like mine. AI’s role is purely accidental, merely a new means of producing content with even less effort than before. AI’s involvement is even implicitly hidden, insofar as most faceless youtubers do not openly admit to using AI in their video production. There is still an artist and an audience, and the possibility of an audience is the principal reason for the existence of the slop. It is purely made for the captivation of the audience. The Faceless Youtuber cares nothing for their art whatsoever. This AI slop is purely exoteric.

This all makes sense to me. What does not make sense, however, are those in the growing minority who make AI slop for their own consumption. Those who are fully aware of the artificial origin of the content they’re consuming and enjoy it anyway. As of now, the technology does not quite exist, as demonstrated by the arduous (if not brief) process I undertook to create my own AI slop YouTube video. In the near future, however, there will inevitably be a platform where prompts can seamlessly create videos of any desired length, and with any amount of creativity and entertainment value that a traditional Youtuber could make.

In time, so too will a platform exist where movies, books, and TV shows could be generated in minutes to the exact specifications and desires of the individual consumer. In the absence of more restrictive copyright regulation of AI content, consumers could even create their own versions of content using established intellectual property, especially when it is for their own entertainment rather than the potential of commercial gain.

Consider again the faceless youtuber, who produces slop for an audience of people. There is clearly still the traditional dynamic between the artist and the audience, no matter how little effort or thought the artist put into the actual art itself.

It is often noted that artificial intelligence’s capacity to recreate art will put artists in jeopardy. This worry is doubtless salient, both in a professional and artistic capacity. Screenwriter’s unions, actors, authors, painters, editors, photographers, among myriad other artistic disciplines have not hidden their opposition to the growth of this technology. This is evidence enough, and it does not need retelling here that artists will be among the first to face replacement at the hands of artificial intelligence, if the process has not already begun.

What is not as often explored, however, is the effect that artificial intelligence might have on the consumption of art. What will it be like to consume and enjoy art once one can use AI to produce exactly what they’re looking for, effectively collapsing the distinction between artist and audience?

In the past, it has not been the case that all possible art exists. It is still not the case, although the time is likely coming soon where all possible art will effectively exist, being a prompt away from generation. Because artists and their labor was relied upon to bring art into existence, there was a limit on how much art could be created and proliferated. There were only so many movie studios with so much money that could produce a certain number of movies in any period. Consumers faced several options, but one consumer and another were presented with more or less the same choices for media consumption. This is all to say that because there was a limited number of options available, people ended up choosing the same options, intentionally or unintentionally.

The Death of Shared Experience

Thus, art as a shared experience formed one of the foundations of culture, with films taking an increasingly large role as the 20th Century progressed. Not only were films a shared experience, but they formed a shared knowledge that laid a basis for culture to exist.
Suppose now artificial intelligence technology progresses to the point of being able to produce films that are both indistinguishable from those made traditionally and perfectly suited to the desires of the individual. There would be no reason to watch the same movie as someone else, because the exact movie you’d like to watch is inevitably going to be different from one person to another. Perhaps there would be “blends” where the artificial intelligence would combine the preference profiles of two individuals into the ideal movie for them to watch together. It’s easy to see that other forms of media, like music and television could evolve similarly.

This is precisely the point. There will be no more distinction between an artist and audience if this form of media production is to become the dominant mode of production. Art in its commercial capacity has relied on exotericism, an intended audience that will consume (for a fee) the media that is created. Market pressures, driven by the often-exorbitant costs of media production, forced producers to appeal to as many consumers as possible to recover costs and maximize profit.

The implications of this outcome would be difficult to understate. Where art’s production costs once created the side-effect of aggregation, which allowed for culture and community to develop, individualized art will almost certainly serve the opposite purpose. With each person watching their own perfect content, there won’t be anything to talk about. There would be nothing to criticize or share an affection for. Even if there was, you’d be the only one to ever care about the way the machine was churning out your slop.

The early effects of the erosion of shared experience are already observable with the way social media algorithms are affecting social interaction. Social media algorithms are nothing other than advanced methods of preference discrimination. The preferences of each user are carefully determined and refined by their interaction with social media, and their content curation is individualized to their tastes. Political preferences, beliefs, and ideologies are among the first things to be detected and tailored to the consumer’s preferences. This amounts to disproportionately being shown things that they already agree with. Instead of being confronted with the beliefs and opinions of others, they only encounter reflections of their own preferences. Instead of having a basis of shared experience upon which discourse can occur, people of different factions, with different sets of accepted facts, become increasingly unintelligible to and isolated from one another. It is not that it makes conversation less likely or less pleasant, but it makes communication impossible.

Art as a temporal process will cease to occur

Traditionally produced art in all its forms has cultivated an organic preference profile for consumers and has provided for more than enough material to train both AI and consumers into knowing what a movie, television show, or book is supposed to look like. Directors, actors, writers, editors, among others have already created the base upon which AI will be able to freely iterate. They created the appetite for media, which is responsible for the economic pressure that led to artificial intelligence to fulfill that demand as cheaply as possible. They will be rendered obsolete by their very own body of work. As firms, production companies would be eager to dispense with the costs of labor. They would flip the switch without blinking.

At this point then, there is only the consumer, and the machine used to entertain them. There would be no more struggle, where the artist is pushing forward and dragging along an audience into new places. Succeeding in some place and failing in others, the struggle between making esoterically interesting art for the artist and exoterically entertaining media for the masses would cease. This struggle has driven artistic development thus far. Instead, we would have endless recursive iterations of preferences that are already learned, that are already perfectly suited to you. It is beyond my scope to tell you where art in all its forms is heading, but it is plainly moving in some direction under the regime of the creative.

You can see this over time through evolving tastes and styles, ideas rising to prominence, falling to criticism. Art is a dialectical process between the artist and the audience. Traditional art might be esoterically interesting to the artist producing for themself, but the significant movement occurs when the other, the audience, is forced to make something of it. The conflict pushing everything forward occurs because the artist and the audience are not one and the same. One can imagine that the artist dares the audience to consume and understand their art, while the audience praises and criticizes accordingly. They’re forced to reconcile with one another, and in this process are both changed in ways that cannot occur in isolation.

In our future there is no more artist that confronts the audience (the other) with their work. There is no artist at all. There is only an audience. If differentiation approaches individualization, then there is only you.

It’s easy to argue that the machine cannot be an artist. While exhibiting immense magnitudes of sophistication, the machine is nothing other than an instrument. Analogous to someone playing a musical instrument, we can put the label of the artist on none other than the consumer whose preferences determine the output of the machine. The consumer, then, becomes the artist too. Once this point is reached, art will find its final resting place.

Art will no longer be an encounter with another, but just a mirror pointed back at oneself. The reflection will approach perfection, so there won’t be any need to change anything. The artist, gifted with the ability to express their preferences exoterically, will no longer find any work in a world where everyone can enjoy their preferences esoterically. There wouldn’t exist any need to consider the aesthetic sensibilities of any other person. There isn’t any need for an argument when we’re only talking to ourselves.

It is unclear what will happen to culture or community when everyone is only engaged with their own movies, sitcoms, music, and AI sports teams. This is because art has not only been a stage for a confrontation with the artist, but also a means of engaging with others in the audience. Discussion about movies, books, and music undergirds much of culture and conversation in general. Individualized media would make such conversations impossible, or hopelessly boring at the very least. Audiences, too, are infamous for not knowing what they want. Oftentimes media that is challenging or unconventional is met with harsh criticism only to be celebrated at its release. This is because art is not always about maximizing happiness or utility. There is art that is miserable, beings no utility, yet people consume it anyways. It will be incredibly difficult for AI to understand this. Audiences would never choose to make Breaking Bad for themselves, and an AI would never create it spontaneously. Media that is genuinely new and challenging is so because it seems unpalatable. It is only the artist who forces us to confront things we thought we didn’t want who can drag us forward.

This Model is Inevitable

In the study of economics, firms are forced to approximate their products to the aggregate desire of consumers, because they are unable to perfectly discriminate between consumers of different preferences and it is economically inefficient to produce many differentiated products. This preference aggregation system will cease to function in the coming years, for several key reasons. Firstly, the scale and depth of data collection regarding individual consumer preferences is bridging the gap between the actual consumer and the approximate consumer that firms once had to use as a stand-in. Spotify, YouTube, Meta, and myriad other firms make their money by effectively learning and developing your preference profile. Because perfect information is profitable for producers, they will never stop. Because consumers enjoy having their preferences perfectly suited, they will never actually stand in the way. On its face, this arrangement is a win-win, which is preciously rare in economic thought.

Secondly, the limitations regarding the cost of media production, which are modeled by the relationship between costs of production and differentiation of the product, are being more and more gradually lifted by the advent and assistance of artificial intelligence. In the past, it would be incredibly unprofitable to make a film for one person. The enormous costs of producing nearly anything meant every piece of commercial media had to cast a wide net. In the future, artificial intelligence will lower the cost of media production to the point where differentiation into media designed for smaller and smaller subsections. Eventually, the market will approach perfect individualization. It is conceivable that individually differentiated media will become normalized.

In a simplistic economic model, this is the inevitable outcome. Consumers will choose the products that are the most suited to their preferences. The producer’s newfound ability to produce unlimitedly differentiated media combined with continuously accumulating capacity to discriminate consumer preferences will make it so tailor-made AI content is able to best match those preferences. As long as consumers are simply maximizing their utility, and producers are maximizing their profit, there is no other way for media production to go.

Still, Maybe Not!

This prediction is leaning on a central assumption though, which is that consumers will rather choose perfectly tailored AI slop to purposefully designed traditional media. I am not certain that this will be the case, for several reasons. Firstly, the reason, I think, why people find art beautiful is the awe of knowing another person created such work. For some time now, video effects have enabled us to bring to convincing reality things that are impossible in the real world. One could manufacture a video of someone doing a kickflip down a 25 stair, but who really cares? It is much more moving to see a young child land their first ollie. Boston Dynamics could create baseball players more perfect than any human, but their perfection is exactly what makes them inhuman and uninteresting. Watching the robot hit a homerun into the parking lot might be fun a couple times, like the first several times one generates an image using AI, but nothing has the staying power of watching a real human struggle to hit the ball and just barely make it on base.

A movie is beautiful because of the work that went into it. Real humans wrote it, directed it, and performed as actors. We are in awe not just of the beauty of the work itself, but the fact that real people were able to bring it to reality. I hope that I’m not falling prey to philistine tendencies here, but I would rather see an imperfect movie made by human beings than a perfect movie made by artificial intelligence. If audiences were to watch that perfect movie, I think most would find it terribly boring. No one cheers on a calculator.

Speaking of boredom brings the next question into focus. With preferences settling wherever they land, audiences might be satiated for some time. There will come a time, though, when people grow bored with the preferences they already know. Artificial intelligence, as a technology that is fundamentally reiterative, relies on its base learning to produce output. In this future, though, the only input it is being trained on is output from other AI. This output is then used as an input to train the process of some other AI, and so on. The process will run in circles forever, never going anywhere new. As of now, only humans have exhibited the ability to come up with new things to feed into the machine. Thus, it could very well be that we would need a human artist to drag us from our stagnation.