AI Policy Report
How Universities Are Regulating AI
We scrape and analyze official AI/generative-AI policies from universities as students take the Project Syllabi survey, then compare them side by side to see where a common standard is emerging — and where schools sharply disagree.
Based on 482 universities · last analyzed August 31, 2026
By the numbers
Policy stance across 475 classified schools
Common requirements
These aren't mutually exclusive — a school can require both disclosure and citation, or layer either on top of any stance above — so the percentages don't need to add up to 100%.
Topics mentioned
Specific phrases and provisions found in policy text, not exclusive categories — same caveat as above, and only topics found in at least one policy are shown.
7 scraped schools haven't been classified yet and aren't included in the counts above.
Highlights
What Most Universities Agree On
Nearly all schools studied (430 of 475 classified universities) explicitly defer final decisions about AI use to individual instructors. This means there is no true university-wide prohibition or blanket permission—instead, the rules you'll follow depend on what your syllabus says. Beyond that, there is broad consistency on the core principle: using AI to complete work and passing it off as entirely your own is treated as an academic integrity violation. The disagreement is not whether that's wrong, but how strictly it's enforced and what disclosure or citation you must do if AI use is allowed.
Where Policies Divide
The 475 classified schools split into five distinct camps, and the largest by far is "varies by instructor" (417 schools). The remaining policies cluster into smaller groups with genuinely different approaches:
- Disclosure-focused (15 schools like Columbia and George Washington): AI use is allowed only if you tell your instructor, upfront.
- Prohibited by default (9 schools like UMass and Florida Atlantic): AI is off-limits unless your instructor explicitly permits it in writing.
- Permitted with guardrails (31 schools like Montclair State and Delaware): AI use is generally okay, but you must cite it and verify accuracy.
- Permissive (3 schools like Western Carolina and UNO): AI is treated as a tool to develop skills; use it openly and follow your instructor's course rules.
Additionally, 106 of 475 schools require disclosure to instructors, and 129 require you to cite or attribute AI use—overlapping requirements that signal a push toward transparency.
The Emerging Standard
The practical reality for most students is that your instructor's syllabus, not your university's headline policy, determines what you can do. Within that flexibility, a loose standard is emerging: if you use AI, be transparent about it and cite it. Schools are not converging on "AI is forbidden" or "AI is free to use"—they're converging on "use it responsibly and let your instructor know." This reflects uncertainty about AI's role in education itself; universities are still experimenting, and they're asking faculty to lead that conversation course by course.
What You Should Do
Read your syllabus carefully before using any AI tool in coursework. If your instructor hasn't addressed AI explicitly, ask. The safest default across nearly all schools is to disclose any generative AI use and cite the tool you used—even if your school doesn't mandate it. This protects you and shows intentional, ethical engagement with the technology.
Look up or compare schools
Search and select one school to read its policy, or select several to compare them side by side.