I like to puzzle. I like to sit at a cold table in South Hall for hours on end and arrange hundreds of pieces of cheap cardboard into garish pictures — pictures I wouldn’t be caught dead hanging on my wall. As someone who enjoys writing, I’ve found that there are many similarities between this process and the process of writing. I like to think that I think about things, so I constantly ask myself why I get so much of a thrill out of such banal activities. As someone who likes to think that they think about things, I realize that my ego loves it; every time I’m right about where a piece fits in, my fragile intellect receives a very small nugget of validation. Puzzling, like writing a paper, consists of these minute ego boosts that build into something more wholly satisfying.
But lately, that cold table in South Hall has grown a little colder, and the most recent puzzle sits unfinished. It seems to me that few people are puzzling; even fewer, I hear, are writing. I’m confronted with this fear on a daily basis, amid depressing conversations with professors, extended discussions with Writing Associates, and résumé review sessions that encourage me to “just put it through ChatGPT.” As a Writing Associate myself, I’ve noticed that fewer people appear to be coming into Oberlin’s Writing Center this semester. It’s hard for us not to immediately attribute this apparent decline to the rapid increase in AI use on college campuses. But, even if appointments are slightly down this semester, there’s very little we can do with this information. While Oberlin certainly has an AI problem, does it follow that AI is the reason that fewer people come into the Writing Center? Common sense says yes. AI is very good at performing certain writing-related tasks and performing them quickly from the comfort of your bedroom, study carrel, or library couch. But AI will make you into a worse writer. Struggling will make you a better one.
Don’t get me wrong: I hate writing papers much more than I hate puzzling. This can sometimes be down to academic pressure or the paralysis of choosing the right words, but, more often than not, I hate writing because it’s incredibly hard. It’s something our brains are more than capable of doing, but not without constantly running into the next problem — a trick of phrasing, an endlessly shifting argument, the gap between the thoughts in your head and the blinking cursor. The new big problem I’m contending with is the urge to simply copy everything into a bot that will tell me what I’m missing. After all, why does it matter to be a better writer? What do you stand to gain from a skill which can (and likely will), for most purposes, be automated?
For one thing, writing forces you to confront these miniature but significant problems all the time. You don’t know what the next sentence in your argument should be, so you try out one, then another, then another, and, finally, uncover from within you the next step forward. For another thing, no one’s going to listen to someone who sounds like everyone else. We’ve already started to develop a collective sense and distaste for AI writing. We almost intuitively shut off, eyes rolled back in our collective skull, when we identify AI writing in the wild. Writing teaches you both what you think and how to express it. You become not just a better thinker, but someone who can express those thoughts convincingly.
But say you don’t use AI to literally write but only to outline a particularly boring essay or point out the gaps in the structure of a body paragraph. In cases like these, it can theoretically produce something akin to what any individual writer could. This type of AI use, though, is perhaps more insidious and damaging than flagrantly shirking a discussion post. Using AI for these undetectably minor problems is like watching someone else solve a puzzle: The puzzle is getting done, but you don’t get the satisfaction of doing it. All that you’re left with is a kitsch painting blown-up to fit 1,000 pieces of cardboard pulp.
This type of problem — an unconvincing thesis statement or a jumbled paragraph — is precisely the point of writing. Like a puzzle, its sole value lies in its being solved. When you plug a messy thesis into AI, it will do a decent job of filling in the gaps in your argument. What it won’t do is show you how to do that next time. AI might help you get somewhere, but it’s not exactly the place you wanted to be, and it’s not “you” getting there. So, next time you have a problem you think you can’t solve alone, don’t type it into ChatGPT. Solve it on your own.
