Do Colleges Check College Essays for AI? What the Common App and Admissions Offices Actually Do

The Common App does not publish any AI-detection step, and we found no college that says on its own admissions pages that it runs an AI detector on application essays. That is not the same as "nobody checks". The Common App's fraud policy treats submitting "the substantive content or output of an artificial intelligence platform" as your own work as fraud. It also leaves detection to each college's "own internal fraud detection protocols". Colleges set their own AI rules, from Brown (spelling and grammar review only) to Georgia Tech (brainstorming and editing welcome). Admissions officers read every essay.

The short answer

• Does the Common App run an AI detector? It publishes no such step. Its fraud policy says it will not independently investigate information about an applicant, except where its own platform is misused.
• Is submitting an AI-written essay against the rules? Yes. The Common App lists passing off AI output as your own work as fraud, and you certify that your application is your own work when you submit.
• Who decides whether to check? Each college, under its own fraud-detection procedures. Colleges can report suspected fraud to the Common App.
• Do colleges say they use AI detectors? None of the admissions policies we read says so. The University of California says it screens responses and may ask you to show the writing is yours. Its news office says it runs plagiarism checks.
• What AI use is allowed? It depends on the college. Most of the policies below allow spelling and grammar review. Some allow brainstorming. None allows pasting in AI-written text.
• What happens if fraud is found? The Common App can suspend or close your account and tell every college on your list. Colleges name outcomes up to revoking admission.

Everything above comes from the Common App's and the colleges' own pages, fact-checked on 26 September 2026. The sources are listed at the end.

The Common App's fraud policy on AI

The Common App fraud policy (last updated August 2023) covers every applicant who uses the platform. Among the forms of fraud it lists is intentionally misrepresenting as your own original work either another person's thoughts, language or ideas, or "the substantive content or output of an artificial intelligence platform, technology, or algorithm".

The process after a report works like this:

1. A report is made. Any registered user, member college or third party can report suspected fraud through the Common App's form.
2. You are told in writing. The Common App gives written notice to the person accused of fraud. If it finds a violation, it sets out the basis for that decision.
3. You can appeal. You have ten business days from the notice to send a written statement and supporting documents. The Common App reviews them within fourteen business days and then makes a final determination.
4. Possible outcomes. Permanent or temporary suspension, or termination of your Common App account, and disclosure of the fraud to the colleges on your "My Colleges" list.

Separately, the Common App's application affirmation asks you to certify that everything you submit is your own work, factually true and honestly presented. It warns that misrepresentation can lead to admission revocation, expulsion or revoked credit and degrees. The affirmation is about authorship. It applies whether or not any software is used.

Who checks: colleges, not the Common App

The fraud policy draws a clear line. The Common App says it will not independently judge, evaluate, verify or investigate information about an applicant, except to address misuse of its own platform. It also says members are not required to report every instance of alleged fraud, but may do so under their own internal fraud-detection procedures.

In practice, that makes checking a college-by-college matter:

• No college we checked publishes an AI-detector step. Brown, Caltech, Cornell, Georgia Tech, Princeton, Yale and the University of Michigan all publish AI rules for applicants. None of them says it runs essays through AI-detection software.
• Some colleges screen for copying. The University of California's Statement of Application Integrity says it runs "regular screenings to verify the integrity of the responses" and may ask applicants to authenticate their writing. UC's explainer on how applications are evaluated says it runs plagiarism checks on applications. UC uses its own application, not the Common App.
• Some colleges use AI to read essays. This is a different thing from detecting AI. Virginia Tech announced in July 2025 that each essay would be reviewed by one human admissions reader and one AI reviewer. The AI is "designed to confirm, not replace, the human reader's score".

The absence of a published detector step does not mean an AI-written essay is safe. It means the risk comes from people reading your essay and from the certification you sign, not from a score you will ever see.

Can admissions readers tell?

Admissions officers read essays alongside everything else in your file: your activities, your teachers' recommendations, your grades and your other answers. An essay that sounds unlike the rest of the application, or that could have been written by any applicant, stands out for that reason alone. Common App's chief executive made the same point in an August 2026 interview with U.S. News: writing that doesn't fit the applicant tends to be noticeable.

Colleges have also been changing how much weight the essay carries. In 2024, Duke's then dean of admissions told The Duke Chronicle that Duke would stop giving essays a numerical rating, partly because it could no longer assume an essay reflected a student's own writing ability. That was a change in how Duke reads essays, not a detection policy.

There is a caution the other way too. AI detectors are not reliable enough to settle authorship on their own. A widely cited Stanford study, Liang et al. (2023), found that popular detectors "consistently misclassify non-native English writing samples as AI-generated". Formal, careful writing is regularly flagged, and short texts swing the most, which matters for a 650-word essay. If you run your own essay through a checker, treat the result as a reason to reread a passage, not as a verdict.

What AI help colleges allow

Each college writes its own rule. These are the official positions we found, quoted from each college's admissions pages on 26 September 2026:

Two patterns stand out. Every policy allows some mechanical help, such as spelling and grammar, and no policy allows AI-written text in the essay itself. In between, colleges differ: Georgia Tech and Michigan welcome brainstorming and editing help, while Brown and Caltech draw the line much earlier. Caltech and Cornell also name translating an essay written in another language as off-limits, which is easy to miss.

If a college you are applying to is not listed, search its admissions site for "AI" or "generative AI". If you find nothing, apply the strictest rule on your list.

If a college questions your essay

1. Read exactly what is being asked. A college may ask you to confirm authorship or explain how you wrote the essay. The Common App process described above only starts if someone files a fraud report.
2. Answer with your writing history. Drafts with dates, document version history, notes and outlines show how the essay grew. A detector score, including one you ran yourself, is weak evidence either way.
3. Use the appeal window if it applies. If the Common App notifies you of a fraud finding, you have ten business days to send a written statement and supporting documents.
4. Be accurate about any AI help you used. If you used AI for something the college allows, such as grammar review, say so plainly and show the draft you wrote.

Our guide to proving you didn't use AI sets out which kinds of evidence carry weight.

A checklist before early deadlines

Many colleges' Early Decision and Early Action deadlines fall on 1 November, and some on 15 November. Check each college's own date.

1. Know the format. The Common App prompts are unchanged for 2026–2027: there are seven prompts, including "Share an essay on any topic of your choice." The personal essay takes 250 to 650 words, as colleges such as the University of Michigan remind applicants.
2. Read each college's AI rule. Use the table above as a starting point, and follow the strictest rule among the colleges you are applying to.
3. Write the draft yourself. Brainstorming help, where allowed, is different from a generated draft that you then edit. Most of the policies above forbid AI outlines or drafts.
4. Keep your drafts. Save dated versions, for example in Google Docs version history. If a college ever asks, that history is your answer.
5. Read it for voice. Replace any sentence that could appear in anyone's essay with a specific detail only you could write. A sentence-level AI-signal check can help you find the generic passages. It cannot tell you what a college will conclude.
6. Don't run your essay through "humanizer" tools. Rewriting to evade a detector turns an allowed essay into a misrepresentation, and it usually makes the writing worse.

For how our own scores are produced and where they fall short, see our accuracy methodology.

Frequently asked questions

Does the Common App check essays for AI?

The Common App publishes no AI-detection step. Its fraud policy treats submitting AI output as your own work as fraud. It says it will not independently investigate information about applicants except where its own platform is misused, and that colleges may follow their own fraud-detection procedures.

Can colleges tell if you used ChatGPT on your college essay?

No college we checked says it uses AI-detection software on essays. But admissions officers read every essay in the context of your whole application, and many colleges' policies say a generated essay can lead to a withdrawn application or revoked admission. Detectors themselves are unreliable, especially for non-native English writers.

Is it allowed to use AI to brainstorm a college essay?

It depends on the college. Georgia Tech, Michigan, Cornell, Yale and Caltech allow some brainstorming help, within limits. Brown allows AI only for spelling and grammar review. None of the policies we read allows submitting AI-written text.

What happens if the Common App finds AI-written content?

The Common App gives written notice and the basis for its decision, and you can appeal within ten business days. Outcomes can include suspending or closing your account and disclosing the fraud to every college on your "My Colleges" list. Colleges can take their own action too, up to revoking admission.

Can I use Grammarly on my Common App essay?

Every college policy we read allows spelling and grammar review. Caltech names Grammarly and Microsoft Editor for reviewing grammar and spelling, and Brown limits AI to exactly that. Features that rewrite sentences or generate text go further than grammar review, so check each college's rule.

Will an AI checker tell me whether a college will flag my essay?

No. Colleges do not publish an AI-scoring step, so there is no score to predict. An AI-signal checker can show which passages read as generic or machine-like, which helps you revise in your own voice. It cannot tell you how an admissions reader will judge the essay.

Sources and further reading

• Common App: Fraud Policy (last updated August 2023)
• Common App: Application affirmations
• Common App: Announcing the 2026–2027 essay prompts (27 February 2026)
• Common App: 2026–2027 essay prompts
• Brown University: Integrity in the application process
• Caltech: Ethical use of AI, guidelines for fall applicants
• Cornell University: Can prospective students use generative AI to help with their application?
• Georgia Tech: Application review and statement on AI
• University of Michigan: Essay questions
• Princeton University: Helpful tips and Dean Richardson on AI (18 August 2025)
• Yale University: AI policy for applicants
• University of California: Statement of Application Integrity (last updated August 2023) and How UC evaluates applications
• Virginia Tech: Admissions updates FAQ (July 2025)
• Liang et al. (2023): GPT detectors are biased against non-native English writers
• Genutext: How to prove you didn't use AI · AI detector false positives · AI detectors and non-native English writers · Accuracy methodology

Common App is a trademark of The Common Application, Inc.; Grammarly is a trademark of its owner. Genutext is independent and not affiliated with, endorsed by, or connected to the Common App or any college named here. Policies are described as published on each organization's own pages as of 26 September 2026; check the linked pages for current details.