Every College Uses the Common App. Most Ignore Its AI Rules.

Every College Uses the Common App. Most Ignore Its AI Rules.

When you click "submit" on a Common Application, you affirm a single sentence that has quietly become the most consequential AI policy in college admissions:

"I certify that all information submitted in the admission process — including this application and any other supporting materials — is my own work, factually true, and honestly presented."

Then a second document — the Common App Fraud Policy — defines what "your own work" means. Since 2024, that definition has explicitly included AI. Submitting "the substantive content or output of an artificial intelligence platform, technology, or algorithm" as your own original work is fraud under the platform's rules.

So far, so clear. One platform, one rule, one signed certification that more than a million applicants click through each year.

Here is what is not clear: across our survey of 174 university AI policies, only 32 schools — about 18% — reference the Common App or its fraud certification in their own admissions AI policy text. The other 142 schools have either written a different AI policy that never mentions the platform their applicants use to apply, or written nothing at all. The result is a two-layer policy environment where applicants sign one document and are bound by another — and most schools never explain how the two fit together.

What the Common App's certification actually says

There are two distinct pieces of language that govern AI use under the Common Application:

1. The Application Affirmation that every applicant signs at submission:

"I certify that all information submitted in the admission process — including this application and any other supporting materials — is my own work, factually true, and honestly presented."

That sentence does not name AI. It does not have to. The Common App's Fraud Policy, which the affirmation incorporates by reference, defines fraud to explicitly include AI-generated content.

2. The Fraud Policy's AI clause, added in 2024, defines application fraud to include:

"submitting plagiarized essays or other written or oral material, or intentionally misrepresenting as one's own original work: (1) another person's thoughts, language, ideas, expressions, or experiences or (2) the substantive content or output of an artificial intelligence platform, technology, or algorithm."

That second prong — the AI prong — is the rule that bites. It treats AI-generated essay content as a category of plagiarism. The legal structure is identical to old-school ghostwriting: misrepresenting another author's work as your own. The Common App's position is that AI is just another "other author."

The policy is not absolute. Running an essay through AI for spelling, grammar, or early-stage topic suggestions does not, on its own, constitute fraud — the triggering word is "substantive." What the Common App prohibits is using AI to generate the actual content — the experiences, ideas, language, and voice — that your essay is supposed to demonstrate are yours.

Enforcement runs through the Common App itself: investigation, account termination, and a report sent to every school on your list. That last consequence is the one most applicants underestimate. A fraud finding is not contained to one application. It is broadcast.

The two-layer problem

The Common App's rule is a platform-level rule. It applies to every applicant who uses the platform regardless of which school they apply to, and it is binding because applicants sign it.

The school-level AI policy is a different document. Each university writes its own (or doesn't), with its own permission level, disclosure expectations, and enforcement signals. Our policy classification rubric sorts them into L0 silent, L1 permissive, L2 line-edits allowed, L3 brainstorm only, L4 prohibited — but none of these school rules erase the Common App certification. They sit on top of it.

This creates a structural ambiguity for every Common App applicant. You are bound by both:

1. The Common App fraud certification you signed at submission, and
2. The school's own AI policy (if it has one), which may be stricter, looser, or silent.

When the two agree, you are fine. When they disagree, the rule that governs is ambiguous unless one of them explicitly resolves the conflict. Most school policies do not. The few that do — and the language they use — are revealing.

The 32 schools that explicitly reference the Common App

In our 174-school survey, exactly 32 schools mention the Common App or its fraud certification in their own admissions AI policy text. Here they are, with their permission/disclosure/enforcement levels in parentheses:

A few patterns jump out.

Most are silent at the L-level but cite the Common App anyway. Twenty-one of the 32 are L0/D0/E0 — they don't have a school-specific AI permission statement. What earns them a spot on this list is that their admissions-integrity page mentions the Common App, usually in the context of explaining what fraudulent application material looks like. They are deferring to the platform without writing their own additional rule.

Schools with their own strong policies use the Common App as reinforcement, not substitute. Brown, Harvard, Yale, Johns Hopkins, and Michigan all have explicit AI policies at the L3 or L4 level and reference the Common App. The two documents work in series: the platform-level fraud rule supplies the legal framework, and the school's own rule supplies the specific permission tier and consequences.

Harvard's FAQ is the clearest version of this layered architecture. From the Harvard College admissions FAQ, AI-generated essay content "violates the Common Application and Coalition on Scoir Application standards as well as the Harvard College Honor Code." Three layers stacked: the Common App fraud rule, the Coalition's parallel rule, and Harvard's own honor code.

Michigan does the same in shorter form: "The University of Michigan expects all applicants to be familiar with and abide by the Common App Fraud Policy" appears verbatim on its undergraduate admissions page. Michigan's own policy is L4, but the citation is to the platform — the platform is doing the legal work.

Brown's integrity page reproduces the Common App's "substantive content or output of an artificial intelligence platform" definition before adding its own independent rule that "the use of artificial intelligence by an applicant is not permitted under any circumstances in conjunction with application content." Brown's stance is stricter than the platform's, but the Common App language is doing the framing.

The 142 schools that don't mention the platform

The remaining 142 schools in our survey are the more interesting category. Every one accepts the Common App as an application route. Every applicant is signing the Common App fraud certification. Yet none of those schools' own AI policies — whether L4 bans, L2 permissive frameworks, or the silent L0 majority — reference the platform's rule.

This silence has three flavors.

Schools with their own AI policy that simply doesn't mention the platform. Stanford, MIT, the UC system, Georgia Tech, Caltech, Carnegie Mellon, Northwestern, Duke. These institutions have written their own AI policies — sometimes elaborate ones, like Georgia Tech's L2 line-editing framework or the UC system's coordinated D3 attestation — but none tells applicants how their rule interacts with the Common App fraud certification.

Silent L0 schools that also don't mention the platform. About 100 of the 142 are classified L0/D0/E0 — no explicit AI policy of any kind. The applicant arrives, finds nothing about AI on the admissions site, and the only governing rule is the Common App certification. The school is effectively delegating the entire AI question to the platform without saying so.

Schools whose policy is in obvious tension with the Common App. A handful have written L1 or L2 policies that permit a level of AI use — line-level editing, paraphrasing, light substantive help — that sits uncomfortably close to the "substantive content" line the Common App prohibits. None of them resolves the tension explicitly. The applicant is left to figure out whether the school's permission overrides the platform's prohibition — which it does not.

The cumulative effect: the most legally binding document — the Common App certification — is the one applicants are least often warned about by the schools they are applying to.

What this means in practice

If you are applying to college through the Common App, you are bound by the Common App's fraud rule regardless of what any individual school's policy says. The platform-level rule is the floor. Individual schools can raise it (Brown, Harvard, Georgetown, BYU, Michigan, SMU all do) but they cannot lower it for you. A school telling you that line-level AI editing is fine — Georgia Tech, Babson, Fordham, the UC system — does not override the Common App's prohibition on submitting AI-generated substantive content as your own.

In practice, the binding rule for almost every Common App applicant in 2026 is:

• Brainstorming with AI is allowed.
• Spelling and grammar checks with AI are allowed.
• Asking AI to generate the actual content, language, ideas, or experiences in your essay is not allowed — regardless of how permissively your specific school's policy is written.

The school-level policy layers on top: it may add a disclosure requirement (D1, D2), a formal AI attestation (D3), or an outright ban (L4). It may also signal enforcement — soft review (E1), AI screening (E2), or a verification interview (E3). Those add to the floor; they do not subtract from it.

If a school-level policy seems looser than the Common App rule, it is either describing a narrower carve-out within the platform's "non-substantive" zone, or silently in tension with the platform rule. Either way, the Common App rule still governs because you signed it.

The practical implication: read both documents, and follow the stricter standard. If you are unsure whether a particular use of AI crosses the "substantive" line, assume it does. The platform's enforcement mechanism — a fraud finding broadcast to every school on your list — is asymmetric. The downside is catastrophic. The upside is marginal.

Why the schools stay quiet

Three plausible reasons emerge from the dataset.

Most schools haven't written an AI policy at all. Our enforcement-gap finding showed that about 70% of universities are L0/D0/E0 — fully silent across all three policy axes. A school that hasn't written its own AI policy obviously hasn't written one that references the Common App.

Some schools consider the rule out of scope for their own site. The platform-level rule is administered by the platform. A school that views the certification as the Common App's responsibility has institutional reasons to leave it alone. Citing a third-party rule on your own admissions page creates ambiguity about which entity is enforcing what.

Some schools may not want to highlight the rule. A handful of schools have program-level AI policies that are more permissive than the Common App fraud rule appears to allow. Drawing applicants' attention to the platform's stricter standard would create a confusing message. Silence is the simpler editorial choice — even if it is a poorer one for applicants.

None of these explanations changes the binding force of the certification. Whether or not a school mentions the rule, every Common App applicant has signed it and is bound by it.

The bottom line for applicants

Three rules of thumb close out this analysis.

1. The Common App rule is the floor. It applies to every applicant who submits through the platform, regardless of what any school's own AI policy says. AI for brainstorming and grammar is fine; AI for "substantive content" is fraud.

2. School-level policies layer on top. They may add attestations, disclosure requirements, or outright bans. They cannot subtract from the platform rule. When they appear to be looser, treat that as appearance, not reality.

3. Read both. Follow the stricter. The 32 schools that reference the Common App in their own policy text are doing the work for you. The 142 that don't are leaving you to layer the rules yourself. Either way, the operational instruction is the same: write the substance yourself, document your process, and reserve AI for the tasks the platform has explicitly carved out as acceptable.

The Common App has done more to shape the AI rules of college admissions than any individual university. Most universities just haven't gotten around to saying so.

---

Browse the full directory of 174 school AI policies, see where schools' rules contradict each other across programs, or read about whether colleges actually use AI detectors on application essays. For a deeper look at the formal AI attestations at top schools, see what you are actually signing when you affirm a college application. The T10 schools' specific positions on AI in admissions and the methodology behind our policy classifications round out the picture.