2026 Admissions174 UniversitiesUpdated February 2026
We analyzed the AI admissions policies of 174 American universities and classified each one across three dimensions: what's allowed, what must be disclosed, and how it's enforced.
Have explicit AI policies
52 universities
Require AI disclosure
15 universities
Actively enforce policies
39 universities
Changed policy since Sept 2025
Updated Feb 2026
Key Findings
From our analysis of 174 university AI admissions policies, reviewed September 2025 and February 2026.
Most universities have no AI admissions policy at all
117 of 174 schools have no explicit AI policy, no disclosure requirement, and no enforcement mechanism (classified L0/D0/E0). Despite widespread concern about AI in college applications, the majority of universities have not issued guidance specific to admissions.
Enforcement is relaxing, not tightening
Between our September 2025 and February 2026 reviews, 82 of 86 enforcement-level changes were in the permissive direction. Schools are removing references to AI detection tools and softening compliance language — even as some are tightening what's allowed.
Schools have internal policy contradictions
56 of 174 universities have at least one program whose AI policy differs from the institution-level classification. Applying to "Duke" does not mean one policy — the Law School prohibits AI while undergraduate admissions permits it. Always check program-specific rules.
See: Duke, Georgetown, Michigan, Stanford
The UC system is the only coordinated multi-campus policy
8 of 10 University of California campuses share E2 enforcement (screening tools). No other multi-campus system in our database has this level of coordination. Most UC campuses classify at L2 (line-level editing allowed) with screening tools in place.
UC Davis, UC Irvine, UC San Diego, UC Santa Barbara, University of California, Berkeley, University of California, Merced, University of California, Riverside, University of California, Santa Cruz
The biggest policy shifts came from unexpected places
George Mason University went from L4 to L0. Harvard University went from L0 to L4. Oberlin College went from L4 to L0. These represent the largest classification changes in a single review round.
Not all changes were restrictive — some schools became more permissive, counter to the general trend.
The Full Distribution
Permission Level
L2
Line-level editing allowed
The vast majority have no explicit policy. Among those that do, L2 (line-level editing) is the most common.
Disclosure Requirement
91% of schools don't mention AI disclosure at all.
Enforcement Method
E3
Formal verification required
78% have no stated enforcement mechanism.
Most common classification
L0/D0/E0
—
117
schools (
67
%)
Explore
Search All 174 Schools
Find your school's AI policy with filters for permission level, disclosure, and enforcement.
Program-Specific Guides
Analysis Tools
Methodology & Framework
Find Your School's AI Policy
Use the search and filters above to find schools, or browse the
full A-Z directory below.
Browse All University AI Policies
174 universities with verified AI admissions policies. Click any school to see their full policy details.
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
R
S
T
U
V
W
Y
About This Data
Data sourced from official university websites|Each policy includes source URL and date accessed|Last comprehensive review: February 2026
Can I use ChatGPT for my college application essay?▾
It depends on the school. Of the 174 universities we analyzed, 30% have an explicit AI policy. The most common stance (L2) permits AI for line-level editing — grammar, style, clarity — but not for drafting or generating content. 70% of schools have no AI-specific policy.
See your school's specific policy in our directory above, or check the AMCAS/TMDSAS platform guidance in our medical school guide.
How reliable are AI detection tools used by colleges?▾
Some schools in our database (classified E2) use AI detection or screening tools. These tools have known limitations, including false-positive rates that can flag human-written text as AI-generated. The bigger risk for applicants is not detection software itself, but submitting generic, AI-sounding prose that lacks personal voice — admissions readers notice this regardless of tools.
Our recommendation: focus on writing authentically rather than trying to evade detection. If a school requires disclosure, disclose honestly.
Where does this data come from?▾
All policies are sourced directly from official university admissions websites, application portals, and official communications. We prioritize admissions-authored content over general academic integrity pages. Each policy includes the source URL and date accessed for verification.
My school isn't listed - what should I do?▾
We're working to add more schools to our database. If your school isn't listed:
- Check their admissions website directly for AI policies
- Search for terms like "AI", "ChatGPT", or "artificial intelligence"
- You can request we add your school at support@gradpilot.com
How often is this data updated?▾
Our data has been through two comprehensive reviews: the initial research in September 2025 and a full re-verification in February 2026. We track policy changes over time and flag schools where the L/D/E classification has shifted.
We conduct comprehensive policy scans before major admissions cycles. Between these full reviews, we welcome corrections and updates at support@gradpilot.com.
Each policy card shows when it was last verified and whether the policy changed. Always check the official university admissions website for the most current information.
Notice something that needs updating? Contact us at support@gradpilot.com