6 Schools Ban AI but Don't Check
Of the 22 universities that restrict or ban AI in admissions materials, 6 have no stated enforcement mechanism. They tell applicants not to use AI, then rely entirely on the honor system.
Key Numbers
The schools with rules but no enforcement
These universities have L3 (brainstorming only) or L4 (AI prohibited) policies combined with E0 (no enforcement stated).
| School | Permission | Disclosure | Enforcement |
|---|---|---|---|
| Haverford College | L3 | D0 | E0 |
| Icahn School of Medicine at Mount Sinai | L3 | D0 | E0 |
| Johns Hopkins University | L3 | D3 | E0 |
| Northwestern University | L3 | D0 | E0 |
| Tufts University | L3 | D0 | E0 |
| University of Michigan–Ann Arbor | L4 | D0 | E0 |
Why it matters
An AI ban without enforcement is effectively a suggestion. Students at these 6 schools face a restrictive policy on paper, but the institution has disclosed no method for detecting violations — no AI detection software, no proctored writing samples, no verification process.
This doesn't mean violations go undetected. Admissions readers may notice generic, AI-sounding prose or inconsistencies between an essay and an interview. But there is a meaningful difference between a school that states it uses screening tools (E2) and one that says nothing about enforcement (E0).
The contrast: schools that do enforce
16 restrictive schools (L3 or L4) do state an enforcement mechanism. Most rely on manual review (E1) — readers flagging voice inconsistencies or contacting applicants directly. A smaller number use screening tools (E2) or formal verification like proctored writing samples (E3).
| School | Permission | Enforcement |
|---|---|---|
| Brigham Young University | L4 | E2 |
| Brown University | L4 | E1 |
| Cornell University | L3 | E1 |
| Dartmouth College | L3 | E1 |
| Georgetown University | L4 | E2 |
| Harvard University | L4 | E1 |
| North Carolina State University | L3 | E1 |
| Southern Methodist University | L4 | E1 |
| Swarthmore College | L3 | E1 |
| UC Irvine | L3 | E2 |
| UC San Diego | L3 | E2 |
| University of San Diego | L3 | E1 |
| Villanova University | L4 | E1 |
| Wesleyan University | L4 | E2 |
| William & Mary | L3 | E1 |
| Yale University | L3 | E1 |
The broader enforcement picture
The enforcement gap is not limited to restrictive schools. Across all 174 universities in our database, 78% have no stated enforcement mechanism of any kind. Even among schools with moderate policies (L2), the vast majority rely on the implicit assumption that applicants will follow the rules.
This pattern likely reflects two realities: AI detection technology remains unreliable for short-form admissions essays, and most admissions offices lack the resources to implement systematic screening. The result is a landscape where the stated policy and the practical reality can diverge significantly.
Methodology
Enforcement classifications are based on what each school's admissions website publicly states. E0 means no enforcement mechanism is mentioned — not that the school never catches violations. Schools may have internal processes not reflected in public materials. See our full methodology for classification details.