Artificial Intelligence and the One Percent
Artificial Intelligence (AI) is a term that gets half of the words right. The richest one percent, the techbros, spend more money than you or I will ever have convincing the rest of us the term means exactly what it implies.
![Artificial Intelligence and the One Percent](/content/images/size/w1200/2024/06/firefly_machine_book.jpg)
Artificial Intelligence (AI) is a term that gets half of the words right. The richest one percent, the techbros, spend more money than you or I will ever have convincing the rest of us the term means exactly what it implies. Those in the artistic world—particularly writers—are slightly irked at this misuse of words, perhaps more so than those in other professions given their daily toil in the word mines; however, it's not the lying that causes artists' blood to boil—it's the stealing.
Large language models, such as ChatGPT, and generative image software like DALL-E, have been well reported as containing works not in the public domain, despite claims to the contrary. And if the stealing and the lying aren't enough, the energy required to power Satan's art engine is estimated to be more than that of some countries, destroying the environment and leaving a trail of dead and starving artists in its wake.
On a more existential level—yes, the boiling oceans and dead coral reefs are bad—he real death to mourn is that of creativity itself. The sacred conversation between an artist and those that lay trembling in awe before the art. A human experience, transformed and transmitted by the artist through a struggle with their medium into a form that connects on a spiritual level with the consumer.
It is a swift death to this struggle that AI promises. The Devil's typewriter or Lucifer's paintbrush offer the result free from the effort required to transform experience, and temptation is only human.
Those that have taken the first bite of this artificial apple know that it is in fact delicious. Sweet creative juices run down their face, exquisite creation absent the effort. Only after the sticky mess dries do they realize it was not creativity at all. But by then it is too late. Their mouth is glued shut, their fingers stuck together, and they are unable to create on their own, like a pornography addict unable to perform in the real world, able only to bash the button that spits out, as it happens in the end, not what they wanted at all.
The slope is gradual and not that slippery, but eventually they find themselves at the bottom of it. What at first were subtle and tolerable flaws manifest themselves through iterative feedback loops. Soon the entire creation is so far from what they wanted it is unrecognizable. But the button must be mashed. Creating is a primal act, and the impulse cannot be denied.
There are extremely good reasons to create art. Witness Stephen Fry reading Nick Cave's letter. Nothing can replace the feeling of conjuring something from the abyss, a new universe still wet from its birth. This has always been true, and will continue to be so. Art is a deeply human activity. A verb. It is something done, and sometimes what it creates is beautiful to those outside of the sacred act of creation, in the realm of aesthetics.
In this sense, art can never be devalued, because it is process. An action that can be taken by the last person to survive the nuclear annihilation of the world; by a child in class at the direction of their teacher; by a writer getting paid for the aesthetic fruits of their sacred act. And, like most beautiful things in this world, money tends to corrupt it. Art it is no longer just a rewarding process for an individual person when its aesthetic fruits become a commodity for mass consumption.
Picture a book in front of you. It has no cover, no illustrations. Just words on a page or screen. Did you enjoy it? This answer already sits inside of you. It formed during the reading, and now exists as an opinion you hold. Now imagine additional information about who wrote the book being provided. At this point, what you feel and what you profess to feel are likely to diverge. Someone of a particular group you admire wrote it, so you describe the lukewarm feeling inside as brilliant. Some author you view as problematic wrote it, so you smother the newly born admiration you've been cradling. Our environment shapes our sense of aesthetics, of how we perceive art, and those at all levels of aesthetic commercialization know this. They attempt to align your desires with their incentives, by creating the very desires that you hold. Ideally you never encounter the results of the artistic process without knowing what to think of them beforehand.
Those currently making a living from practicing their art would understandably like to continue doing so. These artists are the one percent of their industry, able to be paid to engage in the life-affirming artistic process as a career, because a large enough group of people find the output of their labour aesthetically pleasing. And while being devoted to the act of creation is necessary to make a career as an artist, it is not sufficient; thousands of other equally talented people missed out on the ephemeral luck that the one percent stumbled in to. It is true that a person is uniquely capable of creating the art that they create, but they are not unique in their ability to create things that other people enjoy.
There are many possible universes where the dice bouncing around in the cup turned up differently, and we as aesthetic consumers were greeted no with fewer options—simply different ones. In one universe, Whose Names are Unknown was released, and the Grapes of Wrath never was; Sanora Babb could be a household name. Unsurprisingly, the public rhetoric from the one percent of creatives is that the pedestals they stand upon are good actually, and seeing them about to be kicked over by unthinking machines is understandably distressing. No one wants to see the fires from the enemy advancing towards their village.
Many of this group seem to have recently come across the Wikipedia page for Luddite, and like to point out that the Luddites weren't against technology, they were against losing their job. Making a living creating art is hard, and it is soul destroying to conflate the artistic process with the aesthetic admiration of its output—when pure grift attracts millions in venture capitalist funding and years of artistic effort and toil get little, if any, human notice.
This is on top of the fact that existing professionally as an artist is doubly hard in the connected world we live in. One public slip-up could untether the Twitter canon and unwittingly blast away a career. It's why the one percent are so quick to homogenate their public personas, distancing themselves from colleagues—not friends, and demonstrably so as they flee—who attract even the faintest whiff of controversy. No, that person I co-created a movie with isn't my friend, we were barely chums! And now comes AI, unable to participate in or be ruled by the dominant discourse, but capable of churning out mountains of—surely not art—but content that some find aesthetically pleasing.
If AI itself can't be controlled, then the people who might find it interesting must be. AI generated content must be opposed on principle, as a religious person must regard any dealing with the devil. A litany of sins get trotted out: it is objectively inferior, it is theft, it destroys the environment, and it rots your very human soul.
It is commonly stated that the "art" produced by AI is clearly inferior, which begs the question, If it is so obviously of poor quality, then where doth the problem lie? What could there possibly be to fear? The answer is that it is easy to imagine a future in which the quality is not abysmal. AI generated images are already winning contests, and the controversy around the use of AI has good people bowing out of good initiatives. While AI produced content may be judged aesthetically superior to the output of the human artistic struggle, the opposite has also occurred, where a photograph taken by a human won an AI generative image contest. We are already at the point where our aesthetic discernment alone cannot match the morals that we hold.
Those who are simply presented with an image without knowing its provenance are not saddled with the burden of deciding whether or not they are allowed to like it. Even editors opposed to AI have unwittingly selected images generated by it to adorn articles and stories, so it cannot be that the results of AI are always aesthetically inferior to those created by humans.
The most popular objection is that these generative models use stolen works, or at least works which were not explicitly granted permission to be used in training sets. This is both true and indefensible. Commercial products deserve to be compensated monetarily. But this is a temporary complaint. It is not hard to imagine a future state in which models are trained only on public domain, opt-in content. In fact, Adobe has already claimed to do just that with their Firefly model, wisely stating that the images are "safe for commercial use".
In this "safe for commercial use" present and future, artists clearly care about the process—the struggle. Consumers care about the end product. Corporations care about the money. If an author takes three years of self-doubt to drip their blood upon the page, and an AI can create something 80% as good instantly, the corporation is going with the AI. Making a living is about money, and the one percent are used to being able to do this with their art. Unfortunately, they are entitled to nothing. They found themselves both talented and lucky, and AI is taking this luck away from everyone all at once, amid much weeping and gnashing of teeth. Lives are being upended and the status quo destroyed. It is undeniably grim.
If we are being honest, which most of us are reticent to be in public, sometimes a machine does a better job. Sometimes behind a closed bedroom door, it is not the fingers of a human that is craved. The word lotharios do protest, for they are the humans creating the words, sliding them in and out of sentences, building towards a climax of their own imagination, thrusting their vision into our personal experience. But the death of the author is real. There is no guaranteed outcome. No guaranteed conversation. Just a limp book in the hand of a stranger.
Fortunately for the corporations, the money will continue to go to them. They can see a future where they won't have to pay pesky artists for the content they want to shove down our throats. Artists complain that consumers won't be getting the products of their struggle, or even the weaker form or "something made by humans", and they are probably right. The self publishing industry is now flooded by machine generated content. Magazines are being overwhelmed by the same, causing submission pauses, or outright closure.
One might think that surely this means humans would still be able to seek out other humans, but alas, this is not always possible via the traditional avenues, particularly for new or less popular authors. Machines can spit out the words faster than humans, burying the flesh typed stories under tombs of variations on a theme. Programs can also be created to review-bomb humans into oblivion, if the actions of untoward other humans weren't already enough of a problem.
What is legal and what is moral are not necessarily the same thing, so it is completely coherent for someone to claim legally generated AI content is unethical, and this view seems to be the mainstream one floating around the artist community on the internet. The blowback surrounding the cover to Sarah J Maas's book, is a case in point.
In this case, the cover was bad because it was made with AI, not because it was stolen. It was depriving a visual artist the pay to which they must surely be entitled. The leaders of this modern purity movement often post on social media decrying the evils of the machines, and naming and shaming those they feel have sinned against true art. It is the modern day equivalent of church elders in the 1950s standing outside movie theaters ensuring the impressionable youth weren't exposed to secular ideas. In the world but not of it, the phrase now flipped so its application is to that which AI hath wrought. This social pressure is undoubtedly withering, especially to those eking out a living from their art. Espousing any other view is tantamount to career suicide, despite whatever beliefs may be personally held. But this is to forget that art—the struggle, the personal act of creation—is not bothered, or affected by AI at all.
After damning all of our souls, AI is coming to annihilate the environment. The claim is that generative models use vast quantities of electricity and fresh water to produce its derivative content, spewing vast amounts of CO2 while doing so. While it is true that running dataservers requires electricity, and that water is used to cool these datacenters, it is hard to pin down the exact usage that AI is responsible for. As the article states, streaming an hour of Netflix requires ~0.8 kWh, while using a generative text model 1000 times requires ~0.047 kWh, and generating 1000 images requires ~2.9 kWh. Another article estimates that large-language models produce less than 0.5% of all CO2 emissions from the information and communications technology sector.
Although a small percentage of a large number can amount to quite a bit in absolute terms, the people advocating for the abolition of generative AI are just as likely to be tweeting about flying to their latest convention, or binge-watching their favourite streaming service, activities that ask more of our environment than generative AI. True art requires suffering, and in this case, that of the environment. If one were to fall completely down the climate catastrophe hole, an excellent way that an individual can help reduce climate change is by reducing the number of humans. And even though the environment is morally agnostic, the most morally bankrupt among us are not celebrating the recent overturning of Roe-v-Wade in the United States as an environmental victory.
Lastly, it is not just the the environment that is at stake—according to those with the most to lose—but the very soul of what it means to be human. Or the souls of very specific humans, as in the recent AI version of the Tom Waits sound-alike copyright case, where OpenAI seems to have trained an AI model on Scarlett Johanson's voice after she expressly said, "no".
Can AI extend humanity, and take it to places it has never been? Most of the creative one percent have no problems with AI being used to design new cancer-fighting drugs, even if it removes the need for the creative benchtop scientist who has spent her life toiling in the lab, tweaking parameters and running experiments. We are all simply consumers in some aspects of our life, paying no heed to the human cost of our desires.
As with every technological upheaval, the initial fears eventually give way to benefits. Photography didn't replace drawing. Cassette tapes and vinyl records are still sought after. Ripley can premier in black and white. There will always be a market for the past, as nostalgia is a particularly strong drug. And as the saying goes, new ideas are not adopted by humanity through well-reasoned arguments, but by the constant churn of the graveyard.
Even if it is in dark hovels forbidden by the government, humans will gather with other humans to share human-made things for the sheer joy of it. Maybe the most human thing of all is to want both halves of a mutually exclusive offer. To hold the artistic process as sacrosanct and free from the taint of machine oil, to demand that others remunerate its cost of creation, while also desiring the AI-designed cancer drugs and the blissful ignorance of the beauty in the image that was artificially, if not intelligently, produced.