“We live in capitalism. Its power seems inescapable. So did the divine right of kings. Any human power can be resisted and changed by human beings. Resistance and change often begin in art, and very often in our art, the art of words.”
—Ursula K. Le Guin
ChatGPT has lit up the West in the last three months evincing delight among enthusiasts of Artificial Intelligence (AI), but great fear and loathing among critics who see its entry into a world already diminished by machines as a further ratcheting downward of what it means to be human.
The program is from San Francisco-based OpenAI’s “large language model generator,” meaning that it can create an extremely wide variety of responses based on a human-entered prompt. It has already gathered over 200 million individual users, many who are just curious about its capacity to produce when prompted, essays, schedules, outlines, legal briefs, presentations, speeches, emails, letters, creative works like short stories and poetry, even entire novels.
Critics of it evoke sci-fi images of anti-human AI programs in movies such as Skynet in “Terminator 2,” Colossus in “The Forbin Project,” or the ominous Hal in “2001: A Space Odyssey,” that they see as cautionary tales applicable to this new program. But those gleeful about AI’s increased qualitative capacity assure us that while ChatGPT represents the most advanced edge of machine-learning, it is not remotely close to an actual Artificial Intelligence such as portrayed in those dystopian films. Rather, it functions more like an extremely powerful sorting algorithm. Like any computer program, it also has several flaws. Its answers are biased to be both more liberal or libertarian, and it may also be subtly sexist or racist.
The algorithm also has frequent “hallucinations” when there isn’t enough data to compile an answer, meaning that it will make something up to fill in the gaps. For example, when we asked it to write author biographies for our editors, it predicted that two of them were dead (we assure you they are not) and other wildly incorrect information.
All of the protagonists in the famous sci-films were possessed of an arrogance that assumed continued human dominance over their machines (Hello, Dr. Frankenstein). However, in all of the fictional scenarios, the motherboards and central processors began a quick and steep learning curve that soon outpaced those of their creators. The Terminator in the second of the film series tells a young John Connor, who will lead resistance against the machines far in the future, that it learns constantly by processing new information as does ChatGPT.
But how different is this from Facebook, Tik-Tok and other such programs? While Facebook’s algorithm was originally controlled by human input, it is now changing itself without human interaction, in fact, beyond human interaction, to manipulate the wants, interests, and desires of users in a most addictive manner. Facebook’s algorithm is transforming itself into a form of AI that cannot even be deciphered by its creators. It, and the other ubiquitous programs, has become so globally important in people’s lives that it cannot, or at least will not be stopped by the parent corporation, government restrictions, or consumer boycotts. Its influence is that powerful over our daily lives, and all of it is produced by an algorithm that is becoming increasingly independent of humans other than the humans being the mechanics (the servo-protein) to keep the machines running the algorithm.
So, is there a straight line between this new AI program and a future where humanity is either eliminated or relegated to a position of complete dependence upon machines and their algorithms? Probably not. But, just probably. But dystopian fears aside, there are immediate effects that come at the further expense of our humanity. Like most new products out of Silicon Valley, the technology seeks to fulfill a need tech designers created that we did not know yet existed. But, there’s a greater flaw at work than the bugs in the code. There is a further cheapening, dulling, and tapering of the value of human expression for the sake of ruthless, capitalist efficiency that the Luddites saw at the beginning of the Industrial Revolution over 200 years ago.
Although we may not end up being servo-protein for the AI overlords, the erosion of what it means to be human has already made us unable to distinguish between ourselves and the machine, between what is false and what is authentic.
What makes this case especially sinister is that OpenAI pivoted from being a non-profit company when it was founded in 2015 to a for-profit model in 2019. This was done for two reasons: Investors needed to start funneling an insane amount of money to support ChatGPT’s gluttonous operating costs (it runs on the world’s fifth fastest supercomputer), and because of this, the investors want comparable amounts of profit in return. These forces want to make sure that ChatGPT won’t just be a toy novelty that can write Shakespeare sonnets about fast food chains; they want to replace the search engine itself.
The Fifth Estate made a similar observation regarding technological innovation 30 years ago when the publication reluctantly acquired their first computer, writing in the Summer 1993 edition, “The capitalist world, defined by the religion of science and technology, has endlessly created what immediately becomes a necessity.” ChatGPT is positioning itself to become the new necessity in the gig economy, as we are no longer micro-sourcing jobs, but thought.
The new underclass of the techno-feudalist society will be the lowly prompt engineer, as the logical conclusion of the gig economy is for workers to exist in a blur of work-play incentivized only by capital. This is already happening. Some feminists, like Johanna Isaacson (whose recent book is reviewed elsewhere in this issue), have recognized how the effort that goes into social media is basically a new form of unpaid housework, yet social pressures continue to force young people into using products like Snapchat or TikTok or else face being ostracized by their peers. The gig economy owns all.
We loved Airbnb without realizing that we’d eventually have to rent out our own extra bedrooms to compensate for our inflated rent and taxes. We ordered delivery after delivery of Grubhub during the Covid pandemic without realizing we’d have to start working on weekends to pay the hidden fees on our restaurant bills. We have Lyft’d and Uber’d our way into becoming the drivers ourselves.
ChatGPT and other creative AI-generators like Mid-journey and DALL-E have come for the last piece of humanity we believed was immune to automation: our expression, our art.
OpenAI software is biased by race, gender, sex, and political beliefs. They have outsourced the most psychologically harmful work to exploited workers in the least developed parts of the world, such as facilities in Kenya. They use whatever computational power they feel is necessary with no second thought about the damage caused to the environment. And, it completely ignores the millions of skilled writers and artists their programs have been “trained” on.
The reason that the 1% love the gig economy is because every social sub-class must feed on the next lower sub-class to survive, until the very bottom-people are hustling three jobs while being driven mad by YouTube videos about the ease of passive income. AI-generated content replacing human-made content will only make this system even more oppressive.
When asked about the philosophy of anarchism, ChatGPT states that many anarchists might find the software useful because it “decentralizes” notions of power, but this is another hallucination. The power is pointed in one direction—downward—and the blur of content made by humans and AI will make it even more difficult for the underclasses to fight against their oppressors.
Like the internet and social media before it, ChatGPT will change the trajectory of entire industries, and exist as a background presence permanently welded into the content of the internet, as well as the media. In a society where notions of truth and fake news are constantly at war with one another, this injection of the pseudo-real further damages our ability to communicate with each other authentically. We were already scrambling to find each other in the dark corridors of the internet, and now AI generators have placed eerie mannequins at every junction and crossroad.
However, we don’t have to be passive in the war for reality. The sci-fi movies mentioned above give an indication of how to respond.
At the end of “The Forbin Project,” Colossus, the computer network now in charge of humanity, tells Prof. Forbin, its creator, in a creepy machine voice, that “freedom is an illusion,” and that “in time you will come to regard me not only with respect and awe, but with love.” Forbin responds, “Never!”
And, each of us can be a John Connor leading the resistance against the machines.
By slipping into the machine of industrial capitalism and becoming another cog, we not only cease to live authentically, but become willing participants in our own destruction. We should heed the challenge from Ursula Le Guin, as the art of the written word is being infiltrated from within.
Jess Flarity frequently writes for the Fifth Estate.
Peter Werbe is a member of the magazine’s editorial collective.