I have a distinct hatred for generative artificial intelligence (AI). As a creative person, who loves the process of writing essays and deeply cares about the environment and humanity of the world, generative AI is one of the worst things you can do with technology. I stand by the idea that technology can be helpful, but it is ultimately making us dumber, with this being especially true for AI. In a class I took last semester about Child Development, one of our extra credit assignments was to take an essay we had previously written for the class, run the prompt through ChatGPT, and critically analyze the differences between the two. However, instead of generating an essay to write up a comparison about, I did the write up on my dislike of and refusal to use AI. My favorite generative system is my own brain, and the following argument is an adaptation of my response to the aforementioned extra credit.
First and foremost, it’s not that hard to tell when something is created by AI. Sure, you can customize the prompt, fine tune the wording, and use a variety of programs, but it ultimately lacks the human aspect. I’ve had students for classes where I am a Teaching Assistant turn in AI-generated essay responses which I can immediately recognize, especially as I had prior examples of their writing. It’s not just the em dash that shouts “ChatGPT,” AI is stiff and loves to list things in threes, citations are rarely correct and it uses weak sources that are detrimental to analytic papers. Furthermore, AI bots have trouble analyzing evidence. Ironically, analysis is the most important part of academic papers. Most empirical articles won’t just straight up say: “The point you’re trying to prove can be proven by x-y-z,” so analysis shows how you think in terms of the ways in which the evidence you choose ties into the point you’re proving. It is important for your own learning and formation of ideas in a broader context.”
Another example: Returning home for this past break meant an uptick in the amount of cable TV that I consume. In joining my family’s channel surfing, I, for the first time, watched that infamous AI Coca-Cola advertisement in full. I had heard tales of the atrocious AI slop they’d created, but I wasn’t motivated to seek it out. Thus, in my vulnerable state as a casual viewer of afternoon television, I was exposed to what I can only describe as disappointing. A multi-billion dollar company as iconic as Coca-Cola has the money to come up with something creative, but instead they chose a train, squirrels, Santa, and snow, all of which looked different from frame to frame. Additionally, this endeavor was ultimately more expensive and painstaking than hiring a team of artists to animate rather than to supervise the prompts and make sure Santa’s hands have a normal-looking number of fingers. In conversations with my pro-ChatGPT friends, they bring up some valid points about using ChatGPT for studying, learning new things, etc. However, AI is never 100% correct as it draws from wider sources and is not particularly discriminate about discerning truth from things that are in the style of the truth.
A while back, I was speaking with a friend about the calculus homework we were both working on. One problem in particular had us both stumped and searching the web for help yielded nothing. I said something along the lines of, “I wish I wouldn’t feel so guilty if I ChatGPT-ed this question.” And my friend responded, “I would rather use ChatGPT than not get a 100% on my homework.” This gave me pause. My friend would rather be academically dishonest than accept failure or attend office hours. Yes, technically googling “answer + steps” is cheating, but it’s more dishonest in the traditional sense of having your friend write a paper for you, or failing to cite or paraphrase sources. Cheating through AI is so normalized that many students don’t even realize what they’re doing is dishonest. To cheat, AI requires no creativity, no deep internet searching, no tracking down someone extremely good (or at least mildly better than you) at a certain subject and paying them $40 for a B-level essay. Of course, I’m not advocating for cheating in any form, but I lose respect anytime someone talks about using AI to generate a study guide when it would be much more beneficial to their learning to create their own study guide. Even when they say it’s “more efficient” or “saving them time,” that just emphasizes the laziness displayed by many prolific AI users.
Ultimately, I am not against computational AI, but rather, generative. My calculator is my lifeline, even though I am a natural science major. As mentioned prior, I do have an aversion to math, and it comforts me to make sure that four plus five is actually equal to nine. In addition, it’s helpful that computational AI can help computers adapt and learn to better serve their purpose. Sure, generative AI does this too, but it must take pre-existing human materials (art, writing, music, etc.) to learn from, often causing issues with plagiarism and stolen works as many artists do not consent to having their works used to train AI models. AI taking jobs (rather than aiding in them) is definitely another aspect people should be concerned about, but most importantly, AI is taking over the most fundamentally human parts of our cultures.
Even before such modern-day world pillars as short form content, sucky pop artists, and hockey yaoi, we as a society have had the same three basic expressions of emotion and communication: art, music, and writing. Now, AI is actively seeking to destroy the environment and these methods of creation at the same time. Sure, eventually it’ll get good enough to be, in theory, indistinguishable from “real” art forms, but there’s one very obvious tell: a lack of humanity.
So I urge you, pause before you ChatGPT that study guide and maybe consider how finding a better study method will be just as, if not more, beneficial. AI can be used for good, but until then, we must make sure it’s not first overriding our humanity.
