Alright: you suffered through two articles on this already. One about the way the Productivity Paradox killed middle America in the demise of blue collar jobs. Another one predicting a similar demise in white collar jobs due to more automation. But surely the arts are the exclusive domain of humans! How could computers possibly create things full of passion, of desire, of humor, of innovation? How could HAL outshine Van Halen? Could Dall-E best Dali?
It turns out that, no, at least in 2023 that’s not a possibility. Not in 2024, either – at least as current trends hold. Maybe in the distant or not so distant future, and then we are in trouble. But we are in a world of trouble already, so is it really going to be that much worse?
What Is Art?
The question is not philosophical, but practical. On one hand, we think of art as the most advanced and complex human activity, because virtually all humans are handicapped in it. We foster those with a talent for the arts, the rare people that combine technical skills with creativity. And that’s very good, because creativity is very rare, and finding someone that wants to put this already rare capability to practical use is unique.
Humanity has benefited from the arts and from artists for as long as we can remember. Our horizons has been expanded, our beliefs made visible, our passions codified – but also our aspirations targeted, our possibilities enhanced.
How could we possibly automated those things we can’t even imagine yet, because someone with a bigger imagination is there to teach them to us?
Baby Don’t Hurt Me
In the first two article of this series, I talked about how the Productivity Paradox created a strange spiral that sucked away at the ability of workers to make a living. This was due to the fact that productivity increases generate the ability to make do with less work, but also that less work creates a downward pressure on salaries. Also, I noted that with white collar workers, the issue was not whether professions could be entirely replaced thanks to automation, but the percentage of the work performed that could be. If 90% of the work is automated, you need only 10% of the hours from humans, which means 90% of the profession is gone, generating the same downward spiral as in the case of the Productivity Paradox.
With art and artists, the same applies. But with one perverse difference: because the vast majority of individuals is not able to comprehend the depth of an artist’s work, they are perfectly able and willing to do with an automated, inferior version that looks and feels largely identical to them. That is, it is precisely the rarity of the artist’s skill that makes them replaceable.
Of Romance Novel Covers, Pop Song Boredom, and Crappy Web Design
We have seen in the article on the Scholarly Slide that it doesn’t take removing an entire field completely to gut its importance to the economy. Just doing away with the bulk and leaving a rump elite factor is enough. This is true for insurance, medicine, finance, and law, but it also true for the arts.
When we look around us, the vast majority of the art generated is bland and banal. It’s book covers for potboilers, web pages for restaurants, songs that are good enough for the elevator, B-movies that will only survive a single weekend in the theater or just be released directly to a commercial-sponsored streaming service.
And that’s good. It’s very good that art starts small, because there is no realistic way to measure the creativity and innovation of an artist by a single work, especially in the near term. Michelangelo made fun of da Vinci, reportedly, for being outmoded; J.S. Bach’s very existence was almost forgotten for a hundred years; van Gogh famously died in poverty, like countless other break-through artists. Only by providing a field of occupation that is wide and forgiving can we enable creators, because our own competence at measuring the extraordinary is miserable.
The public’s ignorance is the reason we need mediocre artists. We need to give a sufficient number of creatively inclined persons the ability to extract a living out of dull product so they can impress and continue creating. Society as a whole depends on it.
Of Dall-E and Dali
All of that is jeopardized when something shows up that is wonderfully mediocre but automated. If we find computerized creators of art, then their inevitable mediocrity is not going to be a problem, because the vast majority of media consumption prefers the mediocre.
You may have noticed that the highlight image for the series is very visibly computer-generated. For people living in the post-2023 era, that’s visible in the text on the piggy bank: instead of saying THE END OF CAPITALISM, it has a strange mashup of the letters A and T in the middle of the last word. That’s because AI generators in 2023 didn’t know how to write text cleanly, just as they didn’t know how to make human hands with exactly four fingers and a thumb.
I should mention the word, yet, because these are definitely just momentary lapses. The problem with AI is not accuracy, but creativity. The problem with creativity, on the other hand, is that it’s as hard to teach to a computer as it is to be recognized by an average human.
Dall-E and similar AT packages can already generate creations that rival incompetent humans. It’s not a coincidence that ChatGPT is used to cheat on school assignments: those were always the domain of the bored and uninterested. Not surprising that Stable Diffusion is used to create illustrations for blog sites like this: even the simple requirement to add attribution is too much work for someone that has no viewers, no ad revenue, no marketing.
At the end of 2023, software packages that create music of any style are appearing. Just as in the case of ChatGPT or Dall-E, text prompts do the trick: a description of the desired outcome connects with the model, an object of art is generated, voila. Sometimes the outcome is wildly wrong - models for generation of software, legal help, medical diagnosis have been shown to all predict the wrong thing. But as the technology matures, this is going to get better.
Because of the nature of current AI, these models will have a hard to impossible time predicting the next wave. We still need humans for that.
Of van Gogh and Starvation
Sadly, we still need humans, but only for the smallest set of things in this particular domain. We need humans (at this point) to show the way in which current patterns will evolve. Current models are indictive, not predictive, and they cannot imagine something outside the boundary of the now.
It is not a coincidence that so many innovative, creative, ground-breaking artists are also mentally unstable or insane. van Gogh famously cut his ear off, but his painting has become the paragon of the tomorrow, no matter when the now. Maybe it’s the artist’s ability to think, as we like to say, outside the box that makes them seem to the average observer, mentally unstable.
Or maybe it’s the variant of mental processes they possess that makes them able to see things in a way completely different from other, mentally acclimated people.
Of the Future of the Arts
In the article on the Productivity Paradox, we saw how an increase in productivity (per se a good thing) automatically causes a descent into low wage starvation. In the article on the Scholarly Slide, we saw how the assumption the paradox is limited to low-skill work is mistaken: highly skilled, highly educated workers suffer from the same problem, only that their denial is deeper and hence their replacement easier.
In this article, we saw that even when replacing competent, creative people is not really possible (yet), it’s the ability to replace most of their output that is key.
The fundamental problem is that replacing the part of their output that is uniquely human is impossible until the replacement is able to entirely do without humans.
In other words, if we can’t make a living for creative people, we will not be able to innovate. Without innovation, humans will stagnate at our current level.
Of the Future
The series of three articles is one big circle. We saw how increased productivity automatically generates a vast, impoverished underclass. We saw how the educated wrongly think of themselves as immune from the Productivity Paradox because there are small parts of their job that cannot be easily automated. And we saw how the creative are forced to give up the vast majority of their sources of income, because the vast majority of the spenders don’t really care about the creative works they consume.
All in all, this spells doom for humanity. The vast majority of us is fated to increased poverty; more and more of the professions that promised security are going to end up being obsoleted; and even the last domains are doomed because those that could work in them have no way to get to work there.
This is a cycle, a vicious cycle. It’s relatively easy to get out of it, by breaking the first of the links in the chain: the ability of investors to keep the proceeds of increased productivity needs to be abolished, they need to be distributed to everyone. What mechanism is used to get there is irrelevant; in the current framework, a drastic increase in profit taxation at all levels is an easy fix that doesn’t require a lot of change.
The money generated from drastically increased taxation on the corporation and investment side should be used for the logical purpose: to reduce the amount of work required to make a living. This could be in the form of mandatory maximum hours, such that anyone working longer than allowed would cost much more; it could be in the form of a larger social services infrastructure that makes living without hours much more bearable. It could of course be in the form of a completely different mechanism not envisioned or popular, yet.
The only big question left after that reassignment of productivity gains is solved is where to direct the reassigned productivity. Should there be a recognition that all life is fundamentally equal and worthy and full of potential, or should particular competence be rewarded? Should everyone get an equal stipend, or should those we somehow deem worthy by abstract criteria get more?
The answer is left as an exercise to the current situation. The only thing I think is guaranteed is that the current system rewards the wrong people at the expense of the right ones and a whole lot of collateral damage.