False positives · a graduate school personal statement
Why your personal statement reads like AI
If a sentence in your personal statement got flagged, check what it is actually claiming. Often the flag lands precisely where the writing has nothing specific behind it.
A personal statement gets flagged less often for how it is polished and more often for where it goes generic. Readers and classifiers both notice the same thing from different angles: a sentence like 'this experience solidified my passion for law' carries no information a stranger could not have written about any applicant. The machine register survives exactly where you had the least specific thing to say. Read your flagged sentences as a map of the thin spots rather than as an accusation, and you know exactly which paragraphs need a real fact added to them.
Why this happens
A personal statement has one job a resume cannot do: show a reader what it was like to be you in a specific situation. When a writer has a real, singular memory, the sentence tends to carry a name, a place, a number, or an exact exchange of words, and that specificity is exactly what a statistical classifier reads as low-probability, human-shaped text. When a writer is summarizing what a role or experience generically involves, with no scene to point to, the sentence collapses into the same safe, average phrasing a model produces by default, because both are drawing on the same pool of generic ways to describe gaining valuable experience. The overlap is not about polish. It is about how much concrete information the sentence is actually carrying.
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 inflated-experience sentence
- A sentence built on a copula-avoidance verb, 'this internship represented my whole commitment to public service,' inflates an ordinary experience into a grand claim without adding a fact. It is a personal-statement cliche long before it is an AI one, and both exist because the sentence has nothing specific to lean on.
- The role-summary paragraph
- Describing what an internship or job generically involves, rather than one thing that happened during it, produces the same safe phrasing a model defaults to, because neither has a scene to draw from.
- The solidified-my-passion close
- Ending a paragraph by naming the feeling it produced, 'this solidified my desire to pursue law,' is taught as a way to tie experience back to purpose. It also reads as filler wherever the paragraph before it did not already earn the claim.
- The unattached abstract noun
- Words like dedication, passion, and resilience, used with no object attached to them, carry almost no information. A reader and a classifier both register the sentence as generic for the same reason: there is nothing under the word.
What actually helps
Go back through the statement and mark every sentence that could describe more than one applicant. For each one, ask what you actually remember: a name, a specific case, a room, a number, an exact thing someone said to you. If you cannot answer, that sentence was thin before anyone flagged it, and filling it in helps the statement regardless of any detector. Keep your outline and earlier drafts, and if a mentor, professor, or advisor read a version and commented on it, keep that too. If the flag came from an admissions office, ask what specifically triggered it. A specific answer from them narrows the problem to something you can actually address.
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 Explainer (SCOAI): stays with the one case worth describing instead of the broad claim that summarizes it away.
My summer internship at a legal aid clinic served as a testament to my commitment to public service. Working alongside experienced attorneys, I learned not only the value of empathy but also the complexity of the legal system. This experience solidified my desire to pursue a career in law, and I have dedicated myself to understanding legal advocacy ever since. My passion for justice continues to drive everything I do.
My summer internship at a legal aid clinic showed me what public service actually looks like up close. I worked alongside experienced attorneys, and together we learned something worth knowing: the legal system is complicated, and empathy isn't soft — it's essential. That summer is the reason I want to practice law. I've spent every semester since trying to understand legal advocacy better, and I plan to keep learning until the work feels natural. Justice is what I'm after, and it's what keeps me going.
Steps worth taking
- Mark the swappable sentences. Go line by line and flag anything that could describe another applicant with the same job title. That list is your actual editing task.
- Attach a fact to each abstract claim. Wherever you find a word like passion or dedication with nothing under it, add the specific case, name, or number that earned it.
- Keep the drafting trail. Outlines, earlier versions, and any comments from someone who read the statement while you were writing it are the clearest evidence of how it came together.
- Ask what specifically triggered the flag. If an office or reader raised a concern, ask which sentence or section drew it. A vague flag is hard to answer; a specific one usually is not.
Questions
Why is my personal statement flagged as AI?
Usually because one or more paragraphs are written in generic, could-be-anyone language. That register overlaps with machine text regardless of who wrote it, which is why the fix is the same either way: add the specific fact that only you have.
How to make a personal statement not sound like AI?
There is no way to guarantee a score. What reliably makes a statement read as your own is replacing generic claims with specific, checkable detail: names, places, numbers, exact things said. That is also just stronger admissions writing on its own terms.
Can your personal statement be AI?
If you wrote it yourself, it is not, regardless of what a score says. A classifier estimates a probability from patterns in the text; it has no access to whether you were the one typing.
Why is my own writing being detected as AI?
Because the detector is scoring the shape of the sentence, not tracking who produced it. Generic phrasing and machine phrasing look alike to a pattern-matcher for the same reason two people can independently reach for the same cliche.
What happens if you get flagged for AI in university?
It varies by institution and by how seriously the flag is treated. In practice, one score from one tool is a thin basis for a formal finding by itself, and a conversation with the student is usually the first step, but that is a general impression, not a documented policy. Check your specific institution's policy rather than assume.
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.