How We Classify University AI Policies

We evaluate university AI admissions policies across three independent dimensions—Permission, Disclosure, and Enforcement—because a single score cannot capture the nuance of how 174 schools actually regulate AI.

Why three dimensions?

Early attempts to rank AI policies on a single permissive-to-restrictive scale kept breaking down. A university can allow AI use while simultaneously requiring a sworn attestation. Another can ban AI outright yet have no mechanism to enforce the ban. Permission, disclosure, and enforcement are genuinely independent axes.

Consider a real example from our data: Stanford University's institution-level policy is L0 / D0 / E0—no explicit policy, no disclosure requirement, no enforcement. But the Graduate School of Business is coded L4 / D0 / E3—AI prohibited, no disclosure (because they forbid it entirely), and formal verification. A single score cannot represent both realities within the same institution.

L
Permission
L0 - L4
How much AI is allowed?
D
Disclosure
D0 - D3
Must you reveal AI use?
E
Enforcement
E0 - E3
How is it checked?

The Three Dimensions

Each dimension is an independent scale. Higher numbers indicate more restrictive postures.

L

Permission

How much AI assistance can you use?

L0
L1
L2
L3
L4
L0
No Explicit PolicyNo explicit policy

The school has no admissions-specific statement on AI use. A general honesty pledge may exist, but nothing addresses AI tools directly.

Princeton has no admissions-specific AI policy beyond the standard honesty pledge.

L1
Permissive / IntegrativeAI use permitted

AI-generated text may be included in application materials as long as the content is accurate and the applicant takes responsibility. Often paired with a disclosure expectation.

Duke University explicitly permits AI tools in applications and encourages thoughtful use.

L2
Line-Level Editing AllowedLine-level editing allowed

AI may paraphrase, suggest line edits, rephrase sentences, and provide style or clarity suggestions. Wholesale drafting is not allowed; the applicant must write the final wording.

Georgia Tech says AI can help refine ideas, but "your ultimate submission should be your own."

L3
Brainstorming OnlyBrainstorming only

AI is allowed for brainstorming, outlining, topic discovery, and basic mechanics like spelling and grammar. No AI rewriting or paraphrasing of sentences.

Yale allows grammar checks and topic suggestions but considers AI-composed content "application fraud."

L4
ProhibitedAI use prohibited

No AI use of any kind for admissions writing. Brainstorming, editing, and generation are all banned. Some schools specify accessibility accommodations as the only exception.

Georgetown states "use of AI tools to complete any portion of the application is prohibited."

D

Disclosure

When and how must you reveal AI use?

D0
D1
D2
D3
D0
No Disclosure RequiredNo disclosure required

No mention of AI disclosure, even if a general "work is your own" affirmation exists.

Princeton does not ask applicants about AI tool usage.

D1
Optional / EncouragedOptional disclosure

The school suggests applicants may note how AI helped, often with a short text box. Not required, but welcomed.

Carleton College encourages applicants to note any AI assistance but does not require it.

D2
Must Disclose AI UseMust disclose AI use

An explicit prompt or checkbox asks what AI did. Applicants must disclose any AI assistance used during the application process.

Swarthmore College requires applicants to describe how AI was used in their materials.

D3
Must Attest No AI UsedMust attest no AI used

Applicants must certify "no AI beyond X" or "no AI at all," with stated consequences for false attestation. Only assigned when the attestation explicitly references AI.

Georgetown undergrad requires an AI-specific attestation with consequences for violations.

E

Enforcement

How does the school verify compliance?

E0
E1
E2
E3
E0
No Enforcement StatedNo enforcement stated

No enforcement mechanism is mentioned in the policy.

Georgia Tech has clear permission guidance but no stated enforcement mechanism.

E1
Soft / Manual ReviewManual review possible

Readers are advised to watch for voice consistency. The school may request clarification if something seems suspicious.

Yale may conduct manual reviews looking for inconsistencies in voice or style.

E2
Screening ToolsUses screening tools

The school may use AI-detection software, forensic checks, request timed writing samples, or ask for additional drafts.

Some schools run applications through AI detection platforms like Turnitin.

E3
Formal VerificationFormal verification required

Proctored or on-site writing is required. Offers may be contingent on producing an in-person writing sample that matches the application.

Stanford GSB requires applicants to produce on-site writing that is compared against submitted essays.

Distribution Across 174 Schools

Real numbers computed from our full dataset. Last updated with the 2026-2027 admissions cycle data.

Permission Levels

L0
70%
L1
1%
L2
16%
L3
8%
L4
5%

Disclosure Requirements

D0
91%
D1
1%
D2
1%
D3
8%

Enforcement Methods

E0
78%
E1
16%
E2
7%
E3
0%
Most Common Combination
L0/D0/E0
117 of 174 schools (67%)
Program-Specific Overrides
69 schools
40% have at least one program with a different policy

How We Collect Data

Sources

  • Official admissions websites and application portals
  • Official communications from admissions offices (blog posts, FAQs, policy pages)
  • Application platform text (Common App supplements, Coalition App, institutional portals)

Source Priority

Not all university pages carry the same weight. We follow a strict hierarchy:

  1. Admissions-authored AI policy statements — explicit, purpose-built guidance
  2. Application portal language — text applicants encounter during submission
  3. Admissions blog posts and FAQs — official but less formal
  4. General academic integrity policies — used only when nothing admissions-specific exists

Provenance

Every policy in our database includes the specific source URLs where the policy language was found, the exact quotes used to determine the classification, and the date the source was last accessed. You can verify any classification by visiting the linked sources on each school's detail page.

Confidence Levels

High
An explicit, admissions-authored AI policy statement exists. The classification is directly supported by quoted language.
Medium
Policy is inferred from general application instructions, honor code language, or indirect references to AI. The intent is clear but not explicitly stated for admissions.
Low
No relevant language was found. The school is classified based on absence of policy (typically L0/D0/E0). This is the honest answer, not a guess.

Program-Specific Policies

Some universities do not have a single AI policy. Instead, individual schools or programs within the university set their own rules. Our data captures these variations as program-specific overrides.

For example, Georgetown University's undergraduate admissions is L4 / D3 / E2 (AI prohibited, attestation required, screening tools). But the Law School is L3 / D0 / E1 (brainstorming only, no disclosure, soft review), and the Business Programs are L2 / D0 / E1 (line-level editing, no disclosure, soft review).

On each school's detail page, the institution-level code appears first, followed by any program-specific overrides. When applying to a specific program, always use the program-level policy if one exists—it supersedes the institution-level classification.

Georgetown University — Policy by Program
Undergraduate Admissions
L4D3E2
Law School
L3D0E1
Graduate Business Programs
L2D0E1

Explore the Data

See how individual schools are classified, or search across all 174 universities.