How Teachers Detect AI in UCAS Personal Statements

Detecting AI in UCAS personal statements has quietly become part of the autumn workload for sixth-form tutors and heads of year. A personal statement is one of the few pieces of writing where a student has an obvious incentive to sound more polished than they are, and tools like ChatGPT and Claude will happily produce a fluent, confident draft in seconds. When a Year 13 who struggles to write a paragraph in class hands you a statement that reads like a prospectus, you need a fair and consistent way to check before you sign it off.

Why Personal Statements Are Different

Most of the writing you assess has a mark scheme behind it. A personal statement does not. It is a piece of persuasive self-presentation, coached over several drafts, shaped by advice from you, from parents, from older siblings, and increasingly from AI. That layering is exactly what makes it hard to judge by eye - a heavily supported statement and an AI-generated one can look remarkably similar on the page.

There is also a genuine stake for the student. A statement that is substantially written by AI is not the student's own account of why they want to study a subject, and it can misrepresent them to an admissions tutor. Universities are alert to it. Helping a student find their own voice before they submit is part of the pastoral job - and that starts with an honest read of what is actually theirs.

How AI Detection Actually Works

AI detection is probabilistic, not a verdict. On a GradeOrbit school account, detection compares a piece of work with that student's own earlier writing on file and returns a score from 0 to 100 for how far it departs from it, with a confidence label and a plain-English explanation of what it noticed. It does not "prove" that a student used AI, and it should never be treated as an accusation on its own. A high score is a strong signal that the draft is worth a proper conversation; it is not evidence of misconduct.

For a personal statement, the score is most useful read alongside everything else you know about the student. You have their coursework, their lesson contributions, their reference in progress, and often earlier drafts of the same statement. A student whose classwork is halting but whose statement scores very high, sitting far from everything else they have written, is telling you something worth exploring gently. Our guide on how AI detection likelihood scores work explains the reasoning behind the number in more detail.

Where False Positives Hide

Personal statements generate more false positives than almost any everyday classwork, and it is essential to understand why before you act on a score.

• Heavy coaching. A statement redrafted five times with a tutor's feedback naturally converges on a formal, structured, "correct" style - the same qualities that raise an AI likelihood score.
• Template phrasing. Students absorb the standard opening moves ("From a young age I have been fascinated by...") from example statements and school guidance, and that borrowed structure can read as machine-generated.
• EAL and high-achieving students. Students writing in an additional language, or very able students who write formally by habit, can score higher without any AI involvement at all.

This is precisely why a high score is a prompt to look, not a conclusion. The most reliable check you have is the student in front of you: a young person who can talk warmly and specifically about why they want to study engineering, and where the interest came from, is giving you far more than any percentage can.

Handling a High Score Fairly

When a personal statement returns a high likelihood score, treat it as the start of a supportive conversation, not the end of an investigation. Ask the student to talk you through their draft - why they chose that opening anecdote, what a particular claim about super-curricular reading actually refers to. AI-generated statements often contain vague or unverifiable detail: a book the student has not read, work experience described in generic terms, a passion that evaporates under one follow-up question.

Frame it as helping them submit something authentic and strong, because a statement in their own voice will always interview better. Keep a short, factual note of what you discussed and agreed.

How GradeOrbit Protects Student Privacy

A personal statement is inherently personal, so privacy matters even more than usual. How GradeOrbit handles student data depends on which account type you use: teacher AI detection runs on GradeOrbit school accounts under a signed Data Processing Agreement (DPA), where students' work is held with their real names so detection can compare a draft against that student's earlier writing on file. Uploaded scans are removed when a run archives (90 days by default, adjustable by the school), and student records follow the school's retention window and are wiped at offboarding. If a student has no earlier work on file, detection skips them and tells you, rather than guessing. By contrast, on an individual teacher account where no school DPA is in place, uploaded student work for marking is redacted client-side in your browser, processed anonymously as Student 1, Student 2, and so on, and never saved to a database. If you also review sixth-form project work, our guide on how teachers use AI detection scores on the A-Level EPQ covers a closely related task.

Before You Sign Off a Statement

When a personal statement does not sound like the student who handed it in, a score showing how far it departs from their earlier writing gives you a consistent place to start. It is read alongside what you already know - their coursework, what they say in lessons, the story they are trying to tell - and the decision about what to do next stays with you.

Detection is included on school accounts; the GradeOrbit for Schools page explains how it works and what is retained.