Sometimes, but rarely through software. Colleges mostly spot an essay that sounds like nobody in particular, and a growing number now treat AI-written text as application fraud, with real consequences. The safe line is that AI can help your teen think and proofread, while the words themselves are their own. Each college's own rule comes first, so check it before the first draft.
What do colleges actually say about AI in application essays?
There is no single rule, and that is the first thing worth knowing. Admissions offices write their own, and many have not written one yet.
Kaplan surveyed admissions officers at 220 of the top US colleges and universities in July and August 2025 and published the results in November 2025. For writing the essay itself, 2 percent of the schools those officers represented allowed generative AI, 30 percent banned it and 68 percent had no policy. For brainstorming, 27 percent allowed it, 4 percent banned it and 69 percent had no policy. That is a survey of selective US colleges, not of all of them, and it counts schools with a stated policy, not how many essays were checked.
Where a college does speak, it is plain. Yale's admissions statement says that submitting "the substantive content or output of an artificial intelligence platform, technology, or algorithm" is application fraud, and that this "may result in admission revocation or expulsion." It also says that using AI to check grammar or spelling, or to get general advice or topic suggestions at the start, is not fraud. Brown's application integrity page goes further: it does not permit AI use in application content, with one exception for spelling and grammar review. For applicants to the Common App colleges, the Common App fraud policy is the document to read, along with each college's own page.
In the UK, UCAS's guide to using AI in your personal statement (page footer dated 2026) says that using AI to write the statement "could be considered cheating by universities and colleges," while using it "to help you make a start, structure, and then check your final work" is acceptable. Applicants declare that the statement hasn't been copied or provided from another source, "including artificial intelligence software." UCAS also says text that is not in the applicant's own words "could get flagged as fraud," which could "mean you don't receive any offers or have offers withdrawn."
Those are four different institutions saying roughly one thing. My reading: the shared line is not "no AI". It is "the words and the ideas are the applicant's".
Can colleges tell if an essay was written by AI?
Not reliably by software, and that is worth knowing in both directions. It means a detector score is not what exposes an essay. It also means a teenager who is innocent can be wrongly flagged.
OpenAI launched its own AI-text classifier in early 2023 and reported that on a test set of English texts it correctly identified 26 percent of AI-written text and wrongly labeled human-written text as AI-written 9 percent of the time. The page now says: "As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy." A Stanford study reported by Stanford HAI in May 2023 found that seven detectors classified more than half (61.22 percent) of 91 TOEFL essays, written by non-native English speakers, as AI-generated. Those are 2023 detectors and one small set of essays, so treat them as a warning about the method, not a score for today's products. The fair summary is that a tool which flags careful writers, and learners of English in particular, is a poor judge of a teenager's essay.
What admissions readers describe is something simpler. In the Kaplan release, one admissions officer put it this way: "The writing oftentimes also doesn't match the writing of other sections." An application is a set of documents written by one person. A polished, generic essay sitting beside a short answer in the teen's own plain voice stands out, with no software involved.
What I See in Schools
The pattern I see most often is not a teenager who pasted in a whole essay. It is a teenager who asked a chatbot for "a strong opening", liked it, and then wrote the rest in its rhythm. The ideas stayed theirs, but the voice drifted, and by the fourth draft nobody could say which sentences were theirs.
This is also where the school counselor or sixth-form tutor earns their keep. A good one reads the essay beside the rest of the file and asks one question: does this sound like the person I know? That is a human check, and no setting on any tool can stand in for it. If you want to understand how little the work itself gives away on any given evening, how to tell if your child used AI, and how weak most of the signs are is worth reading first.
Where is the line between helping and ghostwriting?
I use the same test I use for homework, because an application essay is the highest-stakes homework a teenager will do:
The Homework Test is one question to ask of any piece of homework before AI touches it: if the machine did this part, what would my child lose? If the answer is nothing, let it help. If the answer is the learning, that part stays with your child.
Applied to an essay, the answer is usually clear. The first of these is help. The second is ghostwriting.
- Checking spelling and grammar, asking "what question does a reader still have after this paragraph?", or asking for three different ways to organize a draft the teen has already written.
- Asking for the opening, the story, the reflection or the closing line, or asking it to "make this sound better" so that most sentences are no longer the teen's.
Two caveats. A college may be stricter than the middle ground (Brown's page, above, is). And the rule applies to us as parents as well: heavy editing by an adult is also a way of submitting words that are not the applicant's. That is my view, not a finding from a source, but I would hold yourself to the same test.
What can I do tonight, by age?
Under 14. There is nothing to submit yet, so the goal is the voice. Let them write about real things in their own words, and read it aloud together. A teen with a practiced voice has no use for a ghostwriter later.
14 to 15. School essays are building the voice a college will eventually hear. Introduce the Homework Test now, on ordinary homework. You could say: "If the computer did this part, what would you lose? If it's nothing, fine. If it's the learning, that bit stays yours."
16 to 17, the application years. Ask them to find the written AI rule for every college they are applying to, and keep it in one place. Then ask to read the essay as a reader, not an editor. You could say: "I'm not going to fix it. Tell me which sentences are definitely yours, and I'll tell you which ones sound most like you."
18 and over. They are an adult applicant, and the declaration is theirs. Your job is to be a first reader and a calm presence, not the author. You could say: "Whatever you decide, I'd rather you tell me where you used it than have it surprise us later."
What if my teen has already used AI on their essay?
Do not panic, and do not submit it. Whether or not a college would call the use fraud depends on what the AI did and what that college's rule says, and those are different for a spell-check and a full draft. The practical route is to go back to their own notes and write the piece again in their own words, then compare it with the draft. That is almost always quicker than repairing a draft that no longer sounds like them.
If something has already been submitted and you are unsure whether it crosses a college's line, the people to ask are the school counselor and, if needed, the college's admissions office. Their rule takes precedence over anything in this post, and over anything a chatbot or a forum will tell you.
What to ask the school
Three questions help, and a counselor can usually answer them in a minute:
- Does the school have a written position on AI in college applications and recommendation letters, and have families seen it?
- Which of our teen's colleges have an AI rule, and where do we find it?
- How does the school help students find their own voice in writing, so that it shows in the application?
A school that can answer all three is ahead of most. If yours has not yet spoken to parents about AI, this is the work I do with schools: helping staff decide where the line sits and say it clearly to families. It is worth sending to your head of year or college counseling office. For the wider question of a teen being asked about AI use, what to do if your child is accused of using AI walks through the first conversations, and the homework version of this question sets out the test in full.
I write a Sunday newsletter that keeps more than 44,000 teachers and parents up to date without the panic.
Dan Fitzpatrick is a former secondary teacher and assistant headteacher who has trained more than 150,000 educators on AI. Read Dan's story.
