The Common App AI Policy: What the Application Actually Says

The Common App has no page called “AI policy.” What it has is a fraud policy, last updated August 2023, that lists the output of an AI system among the things an applicant may not pass off as their own; an affirmation every applicant signs on submission certifying that the application is their own work; and a line of help text beside the essay box telling you not to use AI to write it. Together those three are the rule. The rule is about authorship, and it is quoted below in full, with the page each sentence comes from and the day we read it.

The clause

The fraud policy defines fraud as an intent to deceive or mislead a member college about a candidacy, then lists the circumstances in which it occurs. The third item on that list is the one that mentions AI:

Read it in two halves. The first half is the old plagiarism rule: another person’s thoughts, language, ideas, expressions or experiences. The second half extends that rule to a machine, and it is broader than most people assume. It covers the substantive content of an AI system as well as its output, so an essay built on an idea or a structure a model supplied is inside the clause even when every sentence was retyped. The operative verb is misrepresenting as one’s own: the offence is claiming authorship you do not have.

The Student Solutions Center restates the same item in plainer words, in its list of what counts as fraud:

What you certify when you submit

The affirmations are shown once per college, at submission, and your typed signature is sent to that college with the application. The first of them does not mention AI. It does not need to:

The same affirmation goes on to say what a false certification exposes you to, and this is the sentence an admissions office relies on when it rescinds an offer:

So the fraud policy is the Common App’s own enforcement, and the affirmation is each college’s. An essay a model wrote is a false certification to every school on your list, whether or not the Common App ever hears about it.

The help text beside the essay box

The Student Solutions Center article attached to the personal essay is help text rather than policy, and it shows its age: the site lists it as last modified in September 2023, and it still names Bard, a product that no longer exists. It is still the only Common App sentence that tells applicants directly what to do:

What the rule does not say

Nowhere on commonapp.org is there a list of permitted AI uses. Brainstorming, outlining, feedback on a draft, grammar checking and editing are not mentioned, in either direction. The terms of use, updated in October 2025, do not mention AI at all. We looked for a disclosure field, a counselor rule and a definition of “substantive content,” and found none of the three.

That silence has a shape. The clause is written around misrepresentation, so the question it asks of any AI use is whether the result is still your own work in the sense the affirmation means. Text a model produced that you submit is not. An argument or a story a model suggested that you then wrote up is, by the policy’s own words, its substantive content, and presenting it as yours is the thing the clause names. Feedback that tells you where a draft of your own loses the reader, which you then revise in your own words, is not described by the clause at all, and that is the line most colleges that publish their own rules have drawn as well. The guide to what the published policies say walks through it school by school.

What happens after a report

The policy describes a process rather than a detector. Common App does not scan essays; a report arrives from someone it calls a reporting individual, and the policy names counselors, recommenders, school representatives, parents and member colleges among the people who can be one. On a report, the policy says, Common App sends the applicant written notice naming the source, the applicant has ten business days to answer in writing, and Common App reviews the answer within fourteen. It also says that Common App does not itself verify or investigate the contents of an application on a college’s behalf, and that whatever a college does with its own decision is the college’s to make.

If it finds a violation, the consequences are these:

The last clause is the one that matters. The Common App account is replaceable. A finding sent to every college on your list is not, and it arrives on top of whatever each of those colleges decides to do under the affirmation you signed.

The floor, not the ceiling

The Common App is a platform. Its member colleges set their own rules, and the stricter rule governs the essay you send to each. As of October 16, 2026, 41 of the 153 schools in our directory publish an AI rule of their own for applicants. five of those go further than the Common App and prohibit all AI involvement, feedback included: Brown University, Villanova University, Brigham Young University, Claremont McKenna College, and Georgetown University. The rest permit feedback on writing you did yourself and prohibit AI-written text, which is the Common App’s rule restated with the line drawn explicitly.

The remaining 112 publish no AI rule we could verify on their own pages. For every one of them the Common App clause above is still the rule, because it binds the application rather than the college. Silence from a school is not a permission the Common App has not granted.

• Brown University · prohibits all AI use, feedback included
• Villanova University · prohibits all AI use, feedback included
• Brigham Young University · prohibits all AI use, feedback included
• Claremont McKenna College · prohibits all AI use, feedback included
• Georgetown University · prohibits all AI use, feedback included

Every ban in the directory, quoted · How the directory is checked

A practical reading

Three cases cover almost every question we are asked about this clause.

• A model wrote sentences you kept. Prohibited, however many you changed around them. That is output, presented as your own.
• A model gave you the idea or the structure and you wrote the words. Inside the clause: the policy names substantive content, not just text. Whether a college would ever know is a different question from whether you certified truly.
• A model read your draft and told you what was not working, and you rewrote it yourself. Not described by the Common App’s rule. Described, and permitted, by most of the colleges that publish one, and prohibited by five. Check the school, not the platform.

If a college asks whether you used AI, answer it. The Common App has no disclosure field, but a growing number of schools ask on their own forms, and a true answer to a direct question is the whole of what the affirmation requires.

Where this site stands

CollegeEssayPanel is built for the third case. Seven AI reviewers read a draft you wrote and return scores, a verdict and revision directions; none of them writes a sentence of the essay, so the words stay yours and the certification stays true. Before a review starts, the schools you name are checked against the directory, and a school that prohibits all AI involvement stops the review. That is the Common App clause, applied in the one direction it leaves open.