AI is here, and it’s harming future generations (and the present ones)
An English teacher’s thoughts on “The College Essay is Dead” and ChatGPT.
This time last year, I was in the middle of my one-year stint as a high school English teacher. There were many reasons I left the job (the teacher shortage exists for a reason), but undoubtedly one of the most significant ongoing issues I had was plagiarism.
I went into the year idealistic and passionate. I had also been told that it would be important for me to set classroom routines early and stick with them; these kids were still adjusting to being in the school building full-time, and COVID was making a sudden switch back to remote learning an ongoing possibility.
For these reasons, I made deliberate promises to my students that I was going to structure my courses to represent how scholars engage with literature and how everyday adults use English in their lives. I placed a high emphasis on developing writing skills, and I never gave a single test or quiz. It was supposed to reflect real life.
My rationale was that we were already reading the books together as a class (different from when I was a high school student, in which every single book I read for school was assigned as homework) — therefore, I wasn’t interested in quizzing them on whether they had paid attention during the reading. The students were sitting directly in front of me. I could see if they were focusing or not.
Instead, I tried to be deliberate in making meaningful assignments that would require knowledge and critical thinking about the texts we read, and I told my students that they would struggle to be able to do the projects and writing if they didn’t put in effort to engage with the class content. I figured that if they chose to zone out or not open their books sometimes, that choice would be reflected back into their work. How can you write a nuanced essay or make a detailed project about a novel if you weren’t paying attention to major plot points?
I immediately noticed students submitting high-quality work even though I had seen, day after day, a lack of focus that I was constantly trying to correct. And so my struggles with plagiarism began. The first few months, it was as easy as dropping suspiciously-phrased sentences into Google and finding a perfectly matching result. Some students eventually tried to re-phase the wording a little, but usually I could find a match with similar cadences to the writing and the same keywords, and then they were caught.
No amount of conferencing, contacting parents, or detracting from grades seemed to deter the behavior entirely; many of the students who were penalized early still made multiple cheating attempts throughout the year. They didn’t seem to understand how I was able to spot writing that wasn’t theirs. That is, they weren’t yet mature enough as English scholars to be able to perceive nuances in tone, style, cadence, and content. The irony was not lost on me; this is what I was trying to teach them to do.
To be clear, the plagiarism wasn’t only happening with essays. Students were cheating on five-sentence discussion posts and on basic worksheet answers. They copied directly from the internet onto handwritten annotations. When asked to generate their own question about a text we read in class, they plagiarized this too — Literally, dozens of my students consistently found it preferable to use Google for a one-sentence assignment submission rather than spending two minutes to think of an original idea and write it down.
I spent countless, countless hours throughout the year tracking down the sources my students were using to plagiarize, and I did so because I genuinely cared about whether these kids were making real attempts to develop their writing and (more importantly) to do their own critical thinking. But I also spent countless hours planning assignments that were deliberate and meaningful, and hours in our classes trying to guide my students towards understanding why writing essays — or doing any of the other challenging tasks I asked of them — mattered to their lives.
Especially during a pandemic and after a year of remote and hybrid learning, it’s tough to sell the message I was trying to convey: essays matter because they teach you to perceive nuances in the quality of your own writing, because they teach you to structure and rationalize your thoughts, because you learn about and think more deeply about a text when you write about it, and because a text can teach you important lessons about being a person in this world.
I don’t blame any kid who can’t internalize this line of thinking. School is stressful and demanding already, with the incentives built around strategically turning in the right assignments to get the grade a student wants. Googling answers is a split-second decision that saves time and energy. And, waiting for students when the schoolwork is done, apps and digital games are poised to elicit dopamine with the lowest possible amount of effort. The world is stressful, fun things are simple — Who, under these conditions, could be excited to write an essay, or take on any other academic challenge?
In teacher communities across social media (see TikTok and Twitter especially), you will find the phrase “learned helplessness” again and again. Teachers everywhere are noticing it in their students.
I don’t necessarily like this term. I don’t think the students believe themselves to be helpless, and I don’t think that is what they have learned. A large part of my childhood and my entire adulthood has been spent resisting “kids these days” narratives. Younger generations are not innately weaker or less capable; in fact, it is undoubtable that the youngest among us are in the midst of profound global challenges in fraught political atmospheres, while simultaneously being in a constant state of comparison and competition — on social media, in college admissions, when applying for jobs, everywhere.
I think kids haven’t learned helplessness; they’ve learned to conserve their energy, and they have lacked incentives to put it toward schoolwork. Their anxieties about the world, their social life, and their family struggles are all valid. In comparison, learning for the sake of learning just doesn’t seem that important, especially because they’ve been told that all the knowledge they could ever need is at their fingertips. Cutting corners on schoolwork is an obvious and predictable behavioral choice, and one that we as adults need to understand.
I’m writing about the plagiarism I saw last year because artificial intelligence has become rapidly more available in the short time since I left teaching; this week, a writing AI program called ChatGPT was a huge topic across social media and Medium.com. I still care deeply about my students, and about whether they will have opportunities to access meaningful learning and self-growth in the midst of their extremely stressful lives. These new technologies are promising a world that will be easier for them. In actuality, AI has already reached a state that would allow any of these kids to submit work without ever opening a book or thinking at all about what the text means — without ever really learning anything. I have no doubt in my mind that they will use it. In doing so, I believe their futures are at stake.
A few months ago, in October 2022, my Twitter feed was overrun with bizarre, surrealist images posted by some of the most influential people I followed on the platform. There was an air of excitement in these posts; these were the people who had received early access to DALL-E, an artificial intelligence used to generate images and artwork. The results were often beautiful or accurate in concept, but uncanny when I looked closely at the details.
There was joy to be found because the possibilities for creativity were endless and the results typically silly or abstract. It was fun because the technology was good, but not good enough. It didn’t occur to me to be concerned about it. If anything, DALL-E images were being used for profile pictures and desktop backgrounds and possibly YouTube thumbnails. It wasn’t disrupting anything significant.
Months later, the sentiment around AI has changed dramatically. In November of 2022, Lensa AI made it to the #1 spot on the iOS App Store’s “Photo and Video” category, the result of a trend in which people submitted their portraits and the app produced AI-generated images depicting the subject in various artistic styles.
While the AI trend once again took over Twitter, Instagram, and other social media sites, this time, controversies emerged as well.
In terms of the aesthetic appeal of the images made, the Lensa AI app is undoubtedly producing higher quality work than DALL-E; the trouble is, it’s also producing unethical results. Its biases towards thinness and whiteness, and creation of disturbing content, all stem from the root of the controversy — that Lensa was created by trawling through the internet to capture a total of 2.3 billion captioned images used to train the AI. This includes images that were copyrighted, watermarked, or otherwise never intended for commercial use. The artists used to make the app did not opt in and cannot opt out.
It’s troubling not only because this app is now darkly throwing our worst biases and content back at us, but because it’s also destroying the artists used to create it. It takes years of practice to be able to paint a portrait that would be reminiscent of renaissance-era artwork. As such, an artist commissioned to create one should be able to charge accordingly; depending on the artist, you may reasonably be looking to spend tens, hundreds, or thousands of dollars. Lensa AI costs $7.99.
In response to these concerns, Prism Labs, the company that owns the app, said on Twitter: “As cinema didn’t kill theater and accounting software hasn’t eradicated the profession, AI won’t replace artists but can become a great assisting tool.”
This is bullshit.
While I have no doubt that people who can afford it will value an artisanal, man-made product, there is no denying that this technology will obliterate opportunities for most artists to ever see fair compensation for their talents. Why turn to Etsy to commission a custom portrait when you can get the same result for a quarter of the price? Or, eventually, for free?
More importantly than thinking about individuals who could choose not to commission artwork — why would a company pay for rights to an image if a new, original one can be generated? Why pay artists for their logo designs, photography, or merchandise styling when technology will soon allow all this to be created, perfectly tailored to a company’s target demographics, for no cost at all?
It’s not only the visual artists that are in trouble — it’s everyone.
We’ve already seen the rise of self-checkout machines and self-driving cars. We know that Amazon is growing exponentially as American malls die. These new technologies have already resulted in huge shifts in job markets.
AI is next, it is here, and it is ready to replace untold numbers of workers across a huge span of industries. Not only artists or translators, but writers of all types. Accountants, city planners, and secretaries are at risk. Coding isn’t safe either. Neither is medicine. It’s difficult to think of any industry that won’t be shaped by the capabilities of these new technologies.
But wait — you might be thinking — What if this isn’t a bad thing? Won’t AI make our lives easier? Surely, there is value in doctors getting help in diagnosing patients, office workers getting help in formatting an email, and programmers using AI to solve a bug.
Yes, sure. But let’s also acknowledge what has happened throughout history: new technologies often create new problems even as they solve old ones. After the eggbeater was invented in the mid-1800s, angel food cake skyrocketed in popularity; it just so happens that unlike prior popular cake recipes, angel food cake requires the baker to separate egg whites and whisk them arduously. The work was supposed to get easier, but the tasks got harder. This is a pattern that happens time and time again.
For a more contemporary example, consider email. You no longer have to wait days for a letter from a loved one, or keep meticulous paper files on record in your office; you do, however, have to be available at basically any time of day with few options for excuses if you don’t reply. You can’t leave your work at the office anymore.
We already have a society in which US productivity reached an all-time high in 2021 while inflation grew and wages stagnated. We’re seeing a world in which new technological advances are not alleviating work; rather, we’re doing more of it, not getting compensated fairly, and positioning the youngest generations to live without ever receiving what older generations were guaranteed: benefits, retirement, job security, and fair pay.
AI presents a new type of trouble; it not only can replace menial, repetitive tasks, but also work that has traditionally required creativity and dedicated human involvement. As a humanities major, my peers and I used to find comfort in our bleak job prospects by suggesting that, at least, you need humans to do the humanities. AI is obliterating these comforts.
The industrial revolution tells us what we need to know: new technologies bring with them not only changes to the economy, but widespread societal upheaval. New technologies shape the economy, the economy shapes the people who work within it, and society changes. The question is whether it will change for the better, and at what cost.
I was motivated to write this article after reading Stephen Marche’s piece in The Atlantic, “The College Essay is Dead”. I don’t agree with all the claims he makes in the article; notably, he seems to credit the decline in humanities enrollment as an institutional failure to update their content to modern times, without acknowledging the exponential growth of college costs in the US or the immense cultural pressure to get a career in STEM, which typically leads to higher salaries (especially important given the current state of inflation or the massive changes in the job market due to COVID).
He does, however, aptly lay out a challenge that colleges around the globe will face (and also other academic settings, like my English classroom). AI is now widely and freely available for students to use to produce academic essays — and the results are good.
I tried it myself using ChatGPT, the AI program he discusses in the article. I used a prompt that I had given to my own students when I asked them to write essays on Romeo and Juliet. The quality and level of detail was similar to that of my students. Even if I had been suspicious of plagiarism, I never would have found a way to prove that a student didn’t write the essay. It passed in every way I know to look for.
In his article, Marche lays out this issue in practical terms: “Humanities departments judge their undergraduate students on the basis of their essays. They give Ph.D.s on the basis of a dissertation’s composition. What happens when both processes can be significantly automated? Going by my experience as a former Shakespeare professor, I figure it will take 10 years for academia to face this new reality: two years for the students to figure out the tech, three more years for the professors to recognize that students are using the tech, and then five years for university administrators to decide what, if anything, to do about it.”
Were I still a teacher determined not to let my students cheat, here is what I would have to do: Every time I wanted to assign an essay based on a text read in class, students would first be required to write it in class, by hand, only using a physical copy of the book as a resource. I would record their initial attempt, and then let the students make edits in a typed document — they would share the document with me first, and I would need to check their version histories. They would only be allowed to work on the essays in class, with the screens fully accessible for me to view. Any student whose essay changed too much between their earlier and later drafts would be suspicious — perhaps they quickly Googled something or used an AI when I wasn’t looking, or maybe they cheated at home and then adjusted their work at school. Despite my best efforts, someone would figure out how to get away with plagiarism anyway.
A version of this process would have to be done for any project involving written work. As long as there is internet access, there are opportunities for students to search for answers or information that they need, or to simply have an AI do the task for them. I’m not over-exaggerating; I know, because this is exactly what my students were already doing, but they now have a lower risk of getting caught.
It sounds like hell. Teaching in this way would go against every bit of ideology I have about what meaningful education looks like. It’s difficult to imagine myself implementing a system like this while also trying to convince my students that writing an essay — or doing any assignment — is practice at critical thinking that will enrich them as people, and not a twisted punishment designed by English teachers to make reading and writing more unpleasant for those who already struggle.
I want to be clear that I don’t have some great, innate love for writing essays either. I don’t love writing them now. Any type of writing is hard; writing this article is hard. But, I am lucky enough to have been pushed to take on rigorous work, and to have learned that there is an intrinsic reward to be found in the end — something like satisfaction, like catharsis, like self-actualization.
I don’t mean to be too grandiose. But if you’ll allow me to speak from my own experience, here’s one example of how writing essays has shaped my life:
Knowing that I was going to become an English teacher, I enrolled in a Shakespeare course during my second year of college. I had never enjoyed Shakespeare before, and honestly thought that people who claimed to love his plays were either pretentious or pretending.
The class was fine. I read through the required plays, understood the basics of their plots well enough, and was content to know that I’d be able to teach them if I had to. At some point I’d need to figure out how to make Shakespeare interesting for high schoolers, but that was a problem for later.
In my final for the course, I wrote a comparative essay on two film adaptations of Hamlet. I already read through it once as part of the weekly reading assignments. But, by comparing the two versions and looking through the text again, I ultimately experienced Hamlet four times throughout the term, and then I had to synthesize my experiences into writing. By the time I was re-reading the closing scene as I made the last edits to my paper, the play made me cry. I was no longer struggling to grasp the basic events of the plot or to wrap my mind around the alienating language; I understood the characters, the story, the themes — and they were relevant and moving.
The process that I went through to write my final paper allowed me to understand a very difficult text in a way I was not previously capable of doing. I was able to perceive beauty that was not visible to me before, because I was required to really look and understand. I can now honestly tell you that Hamlet is one of the best things I’ve ever read. I can also tell you that learning to read it gave me skills to discern more meaning from other books I’ve read since; it’s a text that has truly changed my life.
I could cite other anecdotes like this — other times when really, seriously working on a project changed the way I see the world, and changed the kind of person I am.
Is everyone going to have this experience for every assignment? No. Sometimes you just have to write an essay and it doesn’t contribute in any significant way to your life. But the more frequently a person takes on academic challenges, the more likely they are to experience deeper, more personal learning. In the best possible scenario, embarking on these challenges will help a student consider ideas with new perspectives and overcome difficult tasks in the future.
I’m grieving a world where my students, and future generations broadly, are not going to be able to experience the satisfaction and self-growth that school should foster. The new capabilities of technology, the pressure from academic institutions, and the drastically changing job market will create environments with few incentives for critical thinking. Why learn to write anything when AI will do it for us? Why spend years cultivating a particular technique as an artist when AI can reproduce the same style in seconds? Why learn math when AI can do the calculations? Why think critically when another Intelligence has the answers needed to pass the test or do the assignment? Why learn to do any of this if the AI will take over our jobs anyway?
I’m worried that we’re moving toward a future when challenges seem futile and pointless, and therefore, hobbies and skill development seem pointless as well. I’m worried that more and more, we will waste away consuming content but never meaningfully doing our own creative work. I’m worried that this will change our ability to be satisfied with ourselves and our lives, and that we will lose access to beauty and human connection that we currently have. I don’t mean to be over-dramatic — it’s just that I’ve already begun to see it happen.
AI is here to write for us, to make art for us, to do our jobs for us. I don’t have all the answers, or really any answers, about how we should mold our lives and institutions to protect ourselves and future generations from the inevitable downsides that come with these changes. But we need to start thinking about it now, before it’s too late, and before we are too dependent on technology to do the difficult thinking for us.
