False positives · a college application essay
Why your college essay reads like AI
If a counselor or teacher taught you the small-moment opening and the realization turn, you were taught close to the same shape a language model reaches for by default.
A college essay that opens on a small sensory moment and pivots on 'that is when I realized' reads as machine-written because that exact shape is what applications coaching has taught for two decades, and it is also the shape most represented in the essays language models were trained on. The overlap is real. It does not mean your essay is fake. It means the coached structure is a weak signal for authorship, and the fix is to keep the memory and drop the arc that got bolted onto it.
Why this happens
Application coaching converged on one essay shape a long time ago: open on a specific, sensory moment, narrate an obstacle, land on a stated realization, and close by generalizing it into a lesson about who you are. Thousands of workshops, sample-essay books, and college-counseling websites teach this exact arc because it reliably produces a readable 500 words from a seventeen-year-old on deadline. Models trained on large volumes of published personal-essay writing absorbed the same shape, because it is heavily represented in the material available to train on. A classifier built to catch machine text ends up catching this arc, regardless of who wrote it first. Your essay can trip the same pattern-match a generated one would, not because a tool touched it, but because you and the tool learned from the same convention.
Where honest writing and machine writing genuinely overlap
These are real patterns, not coincidences. Each one is something a careful human writer in this situation has good reason to do, and something a model does for entirely different reasons.
- The small-moment opening
- Starting on one concrete sensory detail before naming the topic is standard essay-coaching advice, and it is also the opening move a model reaches for when asked to write a personal essay, because both were trained on the same convention.
- The 'that is when I realized' pivot
- The turn from scene to insight is the single most-taught structural beat in essay workshops. It is also the exact hinge a model uses to convert a memory into a thesis, because the phrase marks that hinge reliably in the material both learned from.
- The closing generalization
- Ending by naming the trait the story proves, 'this taught me resilience,' is coached as a safety net so a reader cannot miss the point. A model reaches for the same closing move for the same reason: it makes the essay legible without asking the reader to infer anything.
- The universal lesson claim
- Framing a personal event as evidence of a broad, portable virtue is standard advice from nearly every application guide. It is also a generic move a model can produce without any real memory behind it, which is why the sentence reads the same whether one exists or not.
What actually helps
Keep what you already have: the earlier drafts in your document history, the annotated version your counselor marked up, notes or a journal entry from around the time of the memory you wrote about. None of that changes a score, but it is real evidence of a process, and it is what a counselor or admissions office actually wants to see if the essay gets questioned. Go to the person who raised the concern and walk them through how the essay came together, in order, from outline to final draft. If a teacher edited a specific paragraph, say which one and why. The goal is not to argue with a number. It is to show the work behind the piece, which a score was never built to see.
What the flattening looks like
On the left, honest writing with the specifics drained out of it, which is the shape that reads as machine-written. On the right, the same content with the writer's own detail restored. The voice is The Wanderer (RCUAI): follows the one specific memory that actually happened instead of building toward a coached turn.
The gym was quiet except for the squeak of my sneakers on the hardwood. As I stared at the scoreboard, I realized that failure was not the end, but the beginning. That is when I understood that basketball had taught me not only discipline but also resilience. Studies show that young athletes who face setbacks develop stronger character, and I am proof of that. From that day forward, I carried this lesson into every part of my life.
The gym was quiet except for the squeak of my sneakers on the hardwood. I stared at the scoreboard, not moving, and it came to me that failure was not the end but the beginning — which is how basketball taught me discipline, and resilience along with it. Other people can point to studies about young athletes and setbacks, I suppose, but I have the thing itself. From that day forward I carried that lesson into everything else, and it's still there.
Steps worth taking
- Find the earlier drafts. Open your document's version history. A dated sequence of drafts, especially with your own comments and edits, is the plainest evidence of how the essay was written.
- Locate the source memory. A journal entry, a text message, or a photo from around the time of the event grounds the essay in something that happened before any draft existed.
- Talk to whoever raised it, in order. Walk the person through the essay from outline to final version rather than defending the finished piece alone. A process is easier to verify than a product.
- Ask your counselor to speak to your drafting process. If a teacher or counselor read early versions, their account of watching the essay change over weeks is evidence you cannot generate after the fact.
Questions
Why is my college essay detected as AI?
Most likely because it follows the small-moment-opening, realization-turn structure that essay coaching has taught for years, which happens to be heavily represented in what language models were trained on. The detector is reading a shape, not a signature.
Why does my essay say 100% AI?
A single score from one tool is not a verdict. These classifiers are pattern-matchers trained on a mix of human and machine text, and a heavily coached, formulaic essay can score the same as a generated one even when no tool touched it.
Can college admissions detect AI?
Admissions offices increasingly run essays through detection tools, but the tools carry real, published false-positive rates. Most offices that use them treat a flag as a reason to ask, not a finding on its own.
What makes an essay sound like AI?
Low variation: sentence lengths that barely change, transitions drawn from a small set of stock phrases, and a structure that resolves too neatly. All three can come from careful coaching just as easily as from a generator.
Do colleges check for AI in application essays?
Some do, using the same commercial tools other institutions use, though there is no shared standard for how a college responds to a flag. If you are unsure whether yours does, ask your counselor or the admissions office directly.
underslop edits drafts for voice. It does not check, score, or contest anything, and nothing on this page will change a result you have already been given. What it can do is show you which patterns in a piece of writing read as machine-written, using the same open lint the rest of the site runs on.