May 3, 2023 (from the archives)
Chat GPT is gaining massive hype as an AI generative output system that can answer any question you could possibly ask it, no matter how specific, peculiar, or outlandish the question. It's convenient and it’s comprehensive. But, it also risks replacing millions of workers and leaving people jobless, propertyless, and lifeless. So, the question must be asked: is Chat GPT an inevitability that advances society, or is Chat GPT the end to all creativity.
While it is a brand new technology, this question is by no means new to our day. Sociologists have been asking this question for hundreds of years each time new tech arises. In fact, if you listen closely to any discussion on Chat GPT you’ll hear Weberian and Marxist terminology from when the industrial revolution and scientific advancements shook their own lifetimes just the same. So, looking at chat GPT through their lenses might prove insightful in our own quest for the future of chat GPT in our own society.
First, there’s the Weberist lens of intellectualization:
Intellectualization is nothing more than the modern day phenomenon of why would I learn that if I can just look it up? Why would I learn that if I can just ask Chat GPT?
We think that this is a new thing, but it was much earlier written about in Weber’s "Sciences as a Vocation." Here, Weber argues that certainty of science directly extinguishes all mystery, magic, and excitement that once existed in the pursuit of knowledge long ago. On an individual scale, people become lifeless and life becomes meaningless as a result of this. He states that life becomes meaningless “since its senseless “progressivity” condemns death to meaninglessness” when he explains,
"Thus the growing process of intellectualization and rationalization does not imply a growing understanding of the conditions under which we live. It means something quite different. It is the knowledge or conviction that if only we wished to understand them we could do so at any time. It means that in principle, then, we are not ruled by mysterious, unpredictable forces, but on the contrary. We can in principle control everything by means of calculation. That in turn means the disenchantment of the world" (13).
Weber’s empirical observation of this phenomenon can be directly applied to Chat GPT, which takes this fear to the extreme as an omniscient, encyclopedic creation with access to the entire internet of human knowledge. Sure, it still gets lots of things wrong in practice, but Chat GPT is nevertheless inching closer and closer to a complete regurgitation of all information across any website spanning the entire internet. At this point, what is the point of learning anything at all?
Weber does not examine the repercussions of this intellectualization, but in a recent panelist discussion, Professor Ben Zhao did. Professor Zhao has the unique background of working directly in the field of cutting-edge AI Tech—even working with a handful of computer scientists that developed Chat GPT! Very telling, Professor Zhao is not optimistic about it. From his perspective as a computer scientist, there is a very real possibility of societal stagnation. According to Professor Zhao, Chat GPT is not anthropomorphic or sentient, like many people have the natural inclination to want to believe. In other words, Chat GPT relies on human input. The problem then arises when it simultaneously discourages human input by mimicking the style of humans’ artwork and as a consequence, it puts poets, song-writers, and artists out of commission. Professor Zhao emphasizes this when he shouted “You are no longer getting commissioned because your name is getting fed into Midjourney thousands of times a month!”
In this way, the algorithm has less and less new human input to go off of, and will incrementally advance at an ever decreasing rate, until one day, it might just remain stagnant. In Professor Zhao's words, “What happens when there are no humans in the industry to generate human inputs? Then we will be stuck with 2025 human data and nothing better.”
In this very Weberian way, Chat GPT has the very real possibility of extinguishing the intrigue and excitement of learning, effectively discouraging new human input, and overall contributing to a stagnant society.
Second, there is the Marxist lens of estranged labor.
Marx fixates on the idea of the estranged labor that he observes during the industrial revolution of his own lifetime. He watched as machines replaced humans in the workforce, making man and machine competitors in a race to higher productivity at lower wages. This competition saw man “reduced to a machine” (286), such that the more man works to keep up with the productivity of machinery, the more machine-like he becomes, and the worse off he is as a human being. This is the phenomenon of estranged labor which states, “the misery of the worker is in inverse proportion to the power and volume of this production” (322).
Estranged labor can be directly applicable to Chat GPT! Just like Marx observes machinery replacing human labor, economists today predict a drastic, possibly catastrophic job replacement by generative AI like Chat GPT—perhaps even more of a job shift than machinery because as a free, universally accessible, and easy-to-use program, Chat GPT can be integrated into the workforce overnight. At least with machines, there was some price and time obstacle to implement the tech into the workforce.
Professor Zhao worries about this. He firmly holds that anything where cost can be optimized at an efficiency of scale will be immediately replaced by Chat GPT or its relative AI algorithms, and facilitate downsizing as well as lowering wages. This process directly encapsulates Marx’s idea of estranged labor: the more that programmers code Chat GPT and the more that humans add content to the input database that Chat GPT relies on for learning, the smarter Chat GPT becomes, and the more jobs it replaces.
Effectively, the more everyone works, the more replaceable they become.
There is a critical—very hopeful—difference in this new age estrangement that differs significantly from the estrangement that Marx describes. Specifically, there is no comparable form of man being depressed to the level of a machine (286) with Chat GPT. This Marxist principle does not apply today because what would it mean for man to be “reduced to a computer?” To start, it is not even a given that this would be a “reduction,” as computers can be quite smart (that’s the whole point of a computer)!
Second, it does not even make sense that man can be reduced to a computer because man physically is incapable of sorting through infinite tables of information and mentally cataloging each topic instantaneously. Third, there is the very widely held view that by letting Chat GPT handle the more menial tasks like indexing, cataloging, and sifting through information that a human would otherwise “waste” time manually doing, the human is more free to be a human. In this way, man does not necessarily become machine-like as Marx describes, but might actually become more human.
This is a very crucial distinction to acknowledge because it could dismantle Marx’s entire theory about the repercussions of estranged labor. To Marx, estranged labor makes man more lifeless, but today, while Chat GPT does facilitate estranged labor, its repercussions do not necessarily make man more lifeless but might just make man more human.
While Chat GPT from a Weberist lens and Marxist lens seems at first to be skeptical of the intellectualization and estranged labor aspects, there’s lots of hope for it after all.
First, to the point of discouraging artists, poets, writers, and other creative jobs that can be “stolen” by Chat GPT and instantaneously mimicked, computer scientists like Professor Zhao are already publishing “computer poison.” Computer poison algorithms distort uploaded images in a subtle way that is not noticeable to the human eye, but to the computer, makes the photo unrecognizable and impossible to be imitated. Poison algorithms like this work for the arts, and can hopefully be applied in their respective way to writing as well. As Professor Zhao profoundly sums up,
“There is hope. We can fight AI with AI!”
Another hopeful fact is the finicky hurdle of copyright that slows Chat GPT use in the meantime. Because AI generated content is deemed “un-copyrightable” because it is not generated by a human, any company with respect for copyright laws will refrain from its use. Lastly, there is hope in halting its “inevitability.” There is the common misconception that this kinda tech is inevitable. This is a lazy, cowardly, excuse to avoid deep critical thinking about its ethics and easily accept it as a given. But, the EU is leading a world-wide push-back to fight this misconception through harsh restrictions, litigation, and demand for chat GPT to at least display its sources. At the more local level, there is also a push-back by education systems that propose regulation for children in schools, and talk of banning its use for students.
Therefore, Chat GPT is a brand new AI generative tech that is shaking our entire world. But, thanks to Weber and Marx, we know that this is nothing actually that new. Using their empirical observations about intellectualization and estranged labor, we can hypothesize that Chat GPT has the frightening potential to be the end of creativity as we know it, replace humans, and create a stagnant society if we let it. But, we don’t have to passively succumb to its supposed “inevitability.” Maybe, if we use it right, with pause, restriction, and critical reflection we can become more human.
With much love,
Arden : )
Insightful as always