As we look around ourselves, more and more of the internet is being spammed with AI content. Human content creators are ringing the alarm bells; their livelihood is under threat, their content which they spent hours making being drowned under a sea of quick-and-easy generators of text, images and videos. Yet, as new as AI is, this phenomenon is old.
The content machine
A large part of what makes a career as a content creator (on YouTube, Twitch, etc) viable is a steady stream of content. To get any sort of stable growth and views, at least a monthly, better weekly upload schedule is recommended, for short-form content like Tiktok, multiple daily uploads are the norm. It is of course possible to operate outside these constraints, but it certainly is most lucrative to have a rather high content output.
The content being produced on these tight schedules is necessarily of somewhat lower quality: No YouTube creator has the ambition to get their videos nominated for the Oscars. Instead, the videos are optimized for other metrics, such as engagement and relatability. For us viewers, this is fine and expected. We don’t really watch a YouTube-video with the expectation that it’ll have us at the edge of our seat like a thriller movie would, instead we as the audience, the viewers merely want to be entertained and content.
We want content that makes us content.
YouTube and TikTok aren’t new in having a bunch of content either. TV shows aren’t generally aiming for oscar-level quality either, with one of the most popular TV formats – reality TV – being infamous for having extremely low production values and artificially and cheaply wrestling the desired emotions in through the use of music, rather than through a compelling story. Compared to movies, TV shows have a lot more content, and make us entertained and content.
And no, movies aren’t immune to this either. Whether it’s unneeded sequels, unneeded prequels, unneeded live action remakes, or just yet another formulaic movie inside a “cinematic universe”, a lot of movies aren’t aiming for the oscars either, but merely for a decent box office return based on a profitable intellectual property or franchise.
Generative AI is just the next step in the content journey. More content that’s not really aiming to do anything at all but to keep us entertained and content enough to continue watching. AI being inferior to artists doesn’t really matter as the entire point of it is to be good enough to achieve a goal in the cheapest way possible. It’s not meant to amaze, it’s not meant to create thought-provoking art, it’s meant to produce content. If movie studios can make a killing of making lazy nostalgia-bait readaptations and sequels and AI is uniquely good at making lazy nostalgia-bait, why should they not use AI for it?
Even if AI replaces all artists, the content machine will continue running undeterred. Fruit Love Island showed as much.
These thoughts aren’t new. Adorno and Horkheimer wrote about pretty much this topic in “The culture industry – Enlightenment as mass deception” in 1944. They even mention the inherent compatibility of content and adverts, a concept so entrenched these days that if YouTube dares to not put ads on a video, creators will cry out for being demonetized and censored.
Capitalism, the killer of art
The problem is, as per usual, capitalism. It’s because creators make money with their YouTube videos, with their TV shows or their movies that they limit themselves so drastically that their art often becomes mere content. If you happen to have made something successful, you better stick to the formula, lest you risk losing subscribers over making something they didn’t expect and didn’t care about.
Creators concerned about AI taking their jobs as creators aren’t chasing creative expression, they’re chasing the success of their content – by wanting to limit how quickly and easily others can create content.
Likewise, creators who are embracing AI also aren’t chasing creative expression, they’re chasing the success of their content – by just making a lot more of it than their competition.
Creators who are chasing creative expression meanwhile see their art plagiarized and ground up into a slop paste that’s liberally and poorly applied to endless productions of content, without a second thought given on what is being expressed, how and why. It’s this sloppy thoughtlessness perhaps which stings most about others using AI. Nevermind the lost monetary gain had people commissioned the creator instead of using AI – art lives from a certain creative dialogue with what came before. Whether it’s inspiration, references, pastiches, quotes or indeed plagiarism, an artist looking to express themselves creatively will have influences from their surrounding world and will, subconsciously or intentionally, let these influences shape their creative expression. Hence the “imitation is the sincerest form of flattery” saying: if someone is imitating you, they’re seeing you as the master.
Generative AI meanwhile doesn’t engage in the same dialog. Instead, it amalgamates everything it can get its hands on and spits out something when prompted. Even if the “prompt engineer” intentionally wanted to plagiarize, he won’t see the artist as the master because he’s not a student. The output is a mere cheap copy, mere content that the AI has generated, which will serve the AI user well as content that ads can be put up on.
If it wasn’t for creators being deeply embedded in the content machine, AI wouldn’t be a threat, but a mere insult and utterly uninteresting, similar to someone taking a picture of the Mona Lisa and now trying to sell it off as his original work. But because creators are so deeply embedded in the content machine, AI now forces them to make a choice: Embrace the content machine and renounce artistry in favor of AI, or embrace artistry and potentially lose their job as a content creator.
AI does threaten jobs
While AI doesn’t threaten art or artist, it does threaten jobs. Right now, this is especially visible for the job of the stock photographer.
Stock photography is already as deeply embedded in the content machine as art can get; not only do professional stock photographers have to pump out immense amounts of images to survive, they also don’t know the context of the whole work their image will be used in. The same photo of a jumping man may be used in an ad for traveling, a newspaper article, a corporate slide deck or a website for diarrhea medicine. As such, any concrete creative vision for the Gesamtkunstwerk is detrimental for the commercial viability of the photo as a stock photo. What does matter for stock photos is craftsmanship and being exactly the kind of content that the user expects.
Users of stock photos generally don’t particularly care about the photo, all they know is that it would be weird if their work didn’t have a visual element of some kind in it. The stock photo mustn’t be too distinct or it’ll force the user to think hard whether it might conflict with the message they want to send. It also doesn’t have to amaze, it merely has to make the user content with using the photo.
This is the ideal environment for generative AI to thrive, it’s practically a perfect match for its capabilities.
Other jobs are at risk right now too, or may get threatened in the future, both outside but especially inside the content machine.
AI causing people to lose jobs or causing creators to no longer profit of their content isn’t a problem per se. Computers (the machines) eliminated the role of computers (the job). This no doubt sucked for the people working as computers at the time, but ultimately, it’s hard to keep a job relevant if a machine can do it cheaper and faster. This isn’t to say that the progression of technology should be lapped up without questioning.
If factories concentrated the means of production to a small, wealthy class, AI (as envisioned by “but imagine in 5 years”-ists) concentrates the act of working itself to an even smaller and even wealthier class of techno-feudalists.
This would be a disaster, not because of the lost jobs, but because of the massive inequality this would cause. The entire genre of cyberpunk is based on this very concept, and it depicts the future as grim. We must not let that happen.
Fighting technofeudalism is out of scope for this blogpost, and I recommend Yanis Varoufakis’ book on the subject for further reading. Let’s go back to art.
What AI lacks is, obviously, the human element. While it can simulate a relationship, it does so in a similar manner to how MDMA would simulate a happy life: It might cause all the right neurons to fire for a while, it might be comforting, but only for as long as you pay the dealer. Even without AI, technological reproductions are just not the same, hence why going to a concert or a museum and seeing works of art in person is special. Indeed, even mundane things like “hanging out” are more special in person than they are on a zoom meeting.
A peek into the crystal ball
It is in this more personal setting where I see the future of art. With AI pushing it out of the role of content for mass media, it would thrive in smaller communities instead – similar to how art operated prior to mass media. Though I’m sure that there’ll be some larger “AI-free zones” and websites on which high-quality human-made content and art will stay relevant on (sites like Nebula come to mind).
As for the big social media platforms, I see them perform similarly to Goethe’s Magician’s Apprentice, initially welcoming the flood of content, but eventually drowning under the weight of petabytes of AI-generated content clogging up every pipeline on their site. There will come a day where YouTube will announce that they’ll auto-delete videos that are old and haven’t been watched in a while, just to buy themselves some breathing room.
Where my crystal ball doesn’t give me a clear picture on is what the effect of content itself will be. For the techno-feudalists, the ideal future would have been keeping us all in relative poverty, and spending every waking hour we’re not creating revenue for them consuming their content – hence why Microsoft, Facebook and Apple all have tried to make VR a thing; the prospect of a virtual reality full of content that they control is way too delicious to pass up. However, being made content by content probably only goes so far, so maybe there’ll eventually be some eating of the rich if the current cyberpunk-dystopia trajectory continues.
I’d also hope that financial support for artists from governments like Creative Europe MEDIA gets widened, liberating artists from constraints of the content machine. It probably is this which artists should be advocating for first and foremost, as it would secure their funding no matter what direction AI ends up developing in.
Leo Wattenberg 