Are you Outsourcing your Thinking?
New York Magazine Thinks So
I was biking and the wind was in my face. Not in my favor though, so my long hair swooshed around my head, brushing my cheeks and obstructing my view through the brown sunglasses. I impatiently swatted it away as I braked to let the pedestrians cross.
Through my headphones, my sister and I talked about the MET gala.
“I didn’t love Emma’s look,” she said.
I nodded, forgetting she couldn’t see me, and tried to recall my Instagram feed from the previous night, sure that I’d seen Emma Chamberlain’s outfit. Before I could ask my sister to remind me what she wore, (a suit-like backless dress), she cut in, her voice already two steps ahead of me:
“You’ll never believe who’s sponsoring her!”
She was referring to Emma’s podcast Anything Goes, where she occasionally interviews guests, gives advice, and tries to make her LA influencer life relatable.
I shook my head, and gave her a few answers that were coherent: online therapy providers, a supplement brand, WebMD…? All were wrong.
“Who?” I asked.
I almost crashed my bike when she said, “ChatGPT.”
The shock was twofold. On one hand, I thought that Open AI was more than well established amidst Gen Z so targeting them, I thought, was slightly redundant. But then, the more evidently shocking one: on Emma Chamberlain’s podcast? Who, as my sister pointed out, has recently been talking about the importance of pushing back against a world that shoves convenience down our throats, who constantly talks about reducing her screen time, and lessening her phone addiction?
Later that day, I tuned into the podcast, and heard that same Emma tell me that through the month of May, while university got particularly hard with exam season, as a student I could use ChatGPT premium for free to help me with my academic needs. I have to give it to her, she didn’t say that it could be used for writing essays, she focused on flashcards and gym schedules, but as someone who’s in the classroom, deep into thesis writing, I know how people are using it.
This is not an anti-ai rant, I think it can be helpful, good even, under certain circumstances, but not when it replaces our ability to engage critically. As a non-chemistry major taking a chemistry class, AI has made understanding certain concepts that I didn’t have a grasp on easier, and slightly smoother. It didn’t become a be all end all solution (these are complicated topics), but rather another tool to use as a cross-checking mechanism. If eventually we can harness the power of LLMs to help us learn and adapt to our ways of thinking, to have a somewhat personalized tutor that helps with understanding and internalizing information, I would be all for it, but when it becomes a tool used from a place of lazy convenience, especially with its large environmental footprint, I start to feel slightly off about it.
This duality shocks me, I know people who are harnessing ChatGPT to help teach themselves coding for their startups, while also knowing people who ask it what movie they should watch, or how to reheat pizza.
A few days ago, New York Magazine released an article on how college students are using AI, and they suggest the use tends more toward the latter. Weaved in through the story of a guy who tried to found a tool to cheat on computing exams during job interviews1, James Walsh writes about a world that if you’re not living in, I’m sure sounds slightly dystopian.
Copying and pasting exact AI responses into a submission box, or thinking that asking it to write a detailed outline with quotes and sources, and then just connecting the dots, somehow doesn’t count as using AI. Professors who are convinced they can spot it from a mile away, but then systematically fail to do so, while students claim that they use ChatGPT to write so that they can spend more time on TikTok. For me, it all culminated in one girl who used it to write a paper with the opening line: “To what extent is schooling hindering students’ cognitive ability to think critically?” and failed to see the irony.
The article is a little more dramatic than what I have observed in Amsterdam. I don’t think any of my peers here believe that pasting exactly what the chatbot gives you is a smart idea, and most are aware that it can make up sources, cite wrong things, and generally not achieve the level of rigor required by a university essay (at least not yet), but still I don’t blame students for their rampant use. If everyone around you is cheating, it’s hard not to do it too, for simple peer pressure purposes, but also for the fear of being left behind. If we really are moving towards a world where LLMs are king, avoiding them all together and failing to understand their strengths and limits is likely not the best approach.
In response to the NY Magazine article, the conversation has turned to designing assignments that are “non Chat-GPT-able.” This has been prevalent at my university since last year, where it was decided that even for “test free classes,” where assignments were usually essays, 50% of all grades now had to be in-class. So, we started answering short form questions, and in class essays. Similarly, while abroad in Australia, for a political economy class, I had to deeply study ten academic papers leading up to the in-class essay, and then use them to bolster my argument once I got the prompt.
These are effective, and though different, they test understanding and synthesis, and while I think all universities will eventually move in this direction, it is a slow and expensive process, creating a strong discrepancy between universities with more stringent AI policies, and looser ones, with implications in the real world that extend beyond having a higher or lower GPA and into generations that graduate with vastly different capacities of critical thinking, analyzing, and interpreting the world around them. A lack of critical thinking is concerning, as it can lead to groups of people who become used to taking everything they read at face-value, and who stop becoming critical of what the “truth” is.
My sister is a junior in High School, so we’ve increasingly been having conversations about university and college, right as my three-year degree comes to an end. If I distill all the International Relations related classes I have taken, I can talk for a long time about theories and concepts that I believe explain what is going on in the world today, but where I think I gained the most, was not in the history based moments, where I learned and regurgitated information, it was not necessarily through the understanding of different political philosophies that shape our current world system, or through analysis of voting patters, incumbent trends, and international courts.
I learned the most as I wrestled with my own essay outlines that didn’t make sense, as I “frankinstined2” my ideas from incoherence to flow, as I synthesized what other writers had said, and critically though about where I could add my voice. I learned the most through the constant practice of turning a jagged rock around until it became smooth; I learned to question, to be patient, and perhaps most importantly to think for myself.
That is my greatest concern with ChatGPT, that it will remove the discomfort of learning, of doing any hard thing really, and continue the acceleration of the convenient and frictionless world we’re building, where thinking becomes outsourced and the easiest answer to find becomes the right one by default. A world where the resilience created by the learning process is lost. Because it’s in the friction, the false starts, the rewrites,3 the uncomfortable task of putting ideas on a page, that the mind is actually shaped. Take that away, and we don’t just lose the process. We lose the very thing the process was meant to build.
This was not very photo oriented but I did write about photography and AI a few weeks ago, you can read it here
I love everything Liam makes, but this reel in particular has kept me thinking.
- ’s recent photoset taken in Uganda. The photos are perfect.
From the Commute:
This Freakonomics mini series on the economics of broadway
One of my favourite podcasts is Search Engine, and last week’s episode did not disappoint
Until Next Week!!
Yes, really
This is what I call the organising and editing phase of an essay after all the ideas are on the page
Also there is something to be said about the bonding power of collective academic suffering, but I’ll save that for another day.