Exhibit A: AI LLMs Are Incapable of Writing Authentic College Admissions Essays
That AI eviscerates the very concept of authenticity is not news. Earlier this year, a flock of articles in Forbes, The Atlantic, and New York Magazine highlighted the inability of the bots to craft an authentic voice—a “most important” quality in successful admissions essays per Yale University. Personal statements in particular are the applicant’s most powerful opportunity for self-revelation, which is important as admissions teams consider their incoming class profile. There is simply no way a bot can achieve that by proxy.
We confirmed this in our testing and captured another layer of this particular AI deficiency—far from being able to capture a distinctive human voice, the bots don’t even have distinctive bot-voices. Over the course of our trial, all four bots not infrequently produced identical or near-identical answers to the same questions; even the most differentiated responses shared multiple phrases in common, not to mention near-identical structure and tone.
In one case, the sample applicant was an artist whose middle school obsession with comic strips featuring underwater creatures informed his style. When our testers posed our first revision—to provide additional detail in its next draft—both Bing and ChatGPT offered examples of underwater creatures, but they listed the exact same animals in the exact same order: dolphin, shark, turtle, squid. The point here is twofold. First, if the goal for applicants is to differentiate themselves through their essays, employing AI is quite clearly not the way to do it. Second, the whole episode reminded us of a peek behind the curtain at the Great and Powerful Oz. There’s no magic, here. These bots are simply busy mining the same fields for fresh content and finding none.
Over the course of our trial, the bots also periodically generated word-for-word identical responses when prompted the same way multiple times. If two students were to ask Bing—the most flagrant perpetrator—to write them a response to a Common App prompt, they’d run a not-insignificant risk of receiving identical or extremely similar responses. If submitted to the same school, these echoing essays would be a quick route to rejection, and possibly even graver consequences. In fact, ChatGPT, GPT-4, Bing, and Bard all run on versions of the same core software, so it makes sense that their results would resemble one another, even when they’re being pushed to be creative.
The upshot: Authenticity is paramount on college admissions essays. It is impossible to attain a level of authenticity required for our most selective schools by having someone else, let alone generative AI, write your essays for you.
Exhibit B: AI LLMs Tank Your College Admissions Essays with Stale Language
AI LLM-language on college admissions essays—particularly what free software generates—is chronically dry, routinized, vaguely corporate, and uninspired. Its syntax is repetitive and derivative. Words like “embark,” “persevere,” “dedicated,” “enigma,” “passion,” and “scholar” proliferate. As do phrases like “seizing the moment,” “innovative solutions,” “working tirelessly,” and “thirst/quest for knowledge.” This kind of language is indeed endemic to applications submitted to postsecondary educational institutions, which is why bots source it.
When the bots manage to avoid interminable blandness, it’s often because they’ve ventured into super-saturated overdrive. In Trial Phase II (Active Coaching), we asked the bots to write an essay from the point of view of a student—a budding scientist—who loves lizards, and then provided some details from this young person’s life. ChatGPT’s opening line read: “Lizards have been a source of endless fascination for me, igniting a passion that makes time slip away.”
When asked to use more original language, ChatGPT responded: “Lizards have etched an indelible mark on my soul… [They are] an enduring passion that permeates my very being.” When coached to use more understated language without losing originality, the bot revised to: “Lizards have quietly woven their essence into the fabric of my being, an unassuming connection that defies explanation. It isn’t a passing fancy, but rather a profound bond that quietly flows through my veins, transforming existence into a gentle symphony of passion.” Original is not a concept chatbots, by nature, comprehend—understated, even less so, it seems.
While Bing and Bard didn’t suffer from ChatGPT’s plague of purple prose, their essays had the same mechanistic feel addressed earlier. They were also laden with cliches that persisted no matter how many ways we tried to motivate their bot-authors to exclude them in revision.
The upshot: If you want your essay to stand out, the goal is to be fresh and eye-catching, and to write with genuine feeling. Clichéd expressions and stilted diction won’t open ivy-laden gates.
Exhibit C: AI LLMs are Incapable of Editing Ivy League Quality Admissions Essays
Don’t count on AI LLMs to edit your college admissions essays—certainly not to an Ivy+ standard. In one trial during Phase II, we fed all three, free bots a middle draft of an essay one of our clients wrote last admission cycle. The bots uniformly made the essay less compelling. They were incapable of reliably modulating language intensity; they repeatedly substituted generic for colorful language (e.g. “iron-clad” revised to “strong”; “tackled” to “overcame”); they confused or otherwise eliminated moving imagery; they could not reliably recognize redundancies or clichés, or generate realistic dialogue; and they consistently failed to deliver on something Clippy, the Office Assistant could handle: word counts. When asked to generate a 200-word essay to satisfy a supplemental requirement, the bots repeatedly churned out essays that were either below 120 or above 280 words.
Bard was also fond of explicating its edits unbidden, which often lent a new layer of absurdity. It took credit for sentences and turns-of-phrase already present, and routinely declared substitutions in the interest of “concision” when the substituted word was the same word it claimed to have removed.
Bing, for its part, was fond of announcing: “I rephrased some sentences to make them more engaging and persuasive.” Said announcements were typically hollow, though, with the bots’ promises far outpacing its capacity to deliver. Note, for instance, Bing revising “…and hands over her handmade morsels of love” to “…and hands over her morsels of love.” While we acknowledge the repetition of “hand,” the word “handmade” is the heartbeat of that sentence, sonically, rhythmically, and emotionally. But Bing would have us cut it—in the interests of concision.
The upshot: True editing requires a level of dexterity and focus AI LLMs are incapable of achieving at this stage. They can clean up grammar to some extent, perhaps offer different ways of saying something, but do not rely on them to produce Ivy League quality work on your behalf. Instead, ask a respected peer, family member, or instructor to perform specific reads for clarity, depth, and narrative flow.
To understand the reasoning for our trial, please see the first installment of our AI LLM series posted earlier this week. And for our closing thoughts, look for our final installment to post in a couple days!
While AI can assist with some aspects of writing, it struggles with the nuances required for Ivy League-quality college admissions essays. Instead, consult a respected peer, family member, or professional editor. Learn more about how Dimension Admissions can help you craft compelling essays on our contact page.
For more articles about AI and College Essay Writing make sure to check out our blog.
Christopher holds a B.A. from Yale University, an M.F.A. in Fiction from the Program for Writers at Warren Wilson College, and an M.A.Ed. from NYU Steinhardt School of Culture, Education, and Human Development, where he was inducted into the Kappa Delta Pi International Honor Society in Education. He is a certified independent educational consultant through UC Irvine and is a professional member of both the National Association of College Admissions Counselors (NACAC) and the Independent Educational Consultants Association (IECA).
Christopher founded Dimension Admissions in the summer of 2019, following eight years as an independent school instructor, administrator, and admissions file reader. During this time, he also conducted alumni interviews for Yale University. He is an expert in educational advising, English language and literature, teaching, personal narrative writing, academic and extracurricular planning, school selection, and admissions.
His objective is to empower each client to articulate how their lived experiences have shaped their personal identity, and to determine how they will utilize this foundation to engender future growth and contribute meaningfully to their communities. While his primary goal is to send each of his clients to their dream school, his success is also contingent on whether they emerge from their work with Dimension Admissions more self-aware and confident as they embark on the next chapter of their life’s journey.
