by Steven Gimbel
In my Philosophy 102 section this semester, midterms were particularly easy to grade because twenty seven of the thirty students handed in slight variants of the same exact answers which were, as I easily verified, descendants of ur-essays generated by ChatGPT. I had gone to great pains in class to distinguish an explication (determining category membership based on a thing’s properties, that is, what it is) from a functional analysis (determining category membership based on a thing’s use, that is, what it does). It was not a distinction their preferred large language model considered and as such when asked to develop an explication of “shoe,” I received the same flawed answer from ninety percent of them. Pointing out this error, half of the faces showed shame and the other half annoyance that I would deprive them of their usual means of “writing” essays.
My comments to them steered away from moralizing about academic integrity and instead I asked what kind of class this was. “Philosophy,” they droned united, sensing the irony. “And what do we do in philosophy?” “Think.” Yes, I told them, in a couple years they would have a boss, likely a romantic partner, maybe even kids, at which point not one of them will care what they think. This may be the last time anyone really does. Why surrender that? Isn’t that what makes you human? Why willingly hand over your very personhood to a machine when this may be your last opportunity to fully embrace it?
I pointed out that never during the semester did I appear in the classroom wearing anything unexpected on my feet. I don’t really need them to tell me what makes something a shoe. The point of the question was not to write down the correct answer. Rather, the value of the exercise was to wrestle with something that seems at first glance trivially easy, but then gets hard when you consider boundary cases. Take this straightforward case and see how tricky it is in order to start building the cognitive muscles you’ll need when thinking about justice, God, truth, or love. It is the process, the struggle, that is important. And that is precisely what our contemporary AI eliminates.
I asked how many work-out and most hands went up. I then asked if they could lift more with a forklift. When they said yes, I asked “Then, why not take one to the gym?” This turned into a utilitarian justification of building skills that will benefit them in their future.
But the fact is that is not the real reason I personally embrace the process. You see, my name is Steve and I am an addict. I started experimenting in high school, became a serious user in college, and now have an everyday habit. In fact, if I don’t read regularly, I go through mood swings and just feel off. Take my books away and I’m a wreck. I think about them incessantly and it is all I can talk about. I am hooked on the thinker’s high.
There is nothing like it when you read a passage that alters your consciousness. The world shrinks. You feel interconnected with reality itself. Your brain congratulates itself with copious amounts of adrenaline, dopamine, and serotonin leaving you floating in ecstasy. We all know the story of Archimedes being so stoned on sudden insight that he ran naked through the streets and what nerd isn’t in search of that level of intoxication?
But that cognitive euphoria requires work. It takes the suffering of bafflement, the frustration of false possibilities, the positing of undiscovered faulty presuppositions. You have to go down dead ends to discover the proper path. But when you do, oh when you do.
Yet it is that initial exasperation the large language models relieve. It seems to be an intellectual pain killer, but it is in fact a cerebral killjoy. These students will never know the glory of the moment of intellectual epiphany.
Aristotle’s genius was pointing out that all things need to be understood as a combination of content and form. ChatGPT strips all texts down to mere content, handing students bullet lists, plot summaries, and charts that streamline the thoughts down to bare information. But it is when we engage the form, the writing, that the reader melds minds with the author and through the intricate twisting of language creates the key to unlock more than the content contains, more than mere knowledge, but insight and wisdom. It is in the engagement with the form, puzzling through passages to the nugget of brilliance that the content fully reveals itself, unfolds inside of your mind creating the preconditions for what Friedrich Nietzsche termed “triumphant self-affirmation.”
But while it may be “self-affirmation,” it becomes even more wonderful when shared. Every intellectual’s romantic dream is to have a partner with whom to share cold, rainy days lolling in bed together, each with a book in the outside hand, inside hands precariously nestled together, held delicately enough to express affection, but loosely enough to give way when a page needs turning. Periodically, each will encounter a well-crafted passage making them stop, reread, and feeling a door in their mind open, experience a Gestalt switch making it so that they bask in the new world they never knew existed but now see in front of them. They will be seized with the irresistible urge to read it aloud, creating an intimate moment. The reader is saying to the partner, “I’m now different and I know you will appreciate this difference and I want you to be different with me because I know you will not only understand how world-changing this is, but I want to be in this other world with you so that we can explore it together.” Such mutual transcendence is the heart of geek fantasy.
Yet that is precisely what the use of ChatGPT disallows. By seeing knowledge as mere facts to be distilled without the struggle that leads to the ecstasy of enlightenment, my students are depriving themselves of one of the most profound delights of humanity and the humanities—not that it is not also to be found in art, social science and STEM fields as well.
I gave them the chance to redo their exams and exasperatedly begged them to really show me their inner-battles with the questions. “No pain, no gain,” I told them framing the payoff in the sort of monetary terms which seem to motivate them. But deep down I knew that a few of them would find something better, the thinker’s high. And when they got a taste of it, I let them know that I will be teaching other classes next semester thereby becoming their regular go-to, their dealer, providing new and increasingly stranger ideas that lead to emotions ChatGPT could never produce.
***
Enjoying the content on 3QD? Help keep us going by donating now.