Research:Wiki Education in the Age of Generative AI: Measurement and Intervention
This page documents a research project in progress.
Information may be incomplete and change as the project progresses.
Please contact the project lead before formally citing or reusing results from this page.
We're exploring the impacts of Large Language Models (LLMs) usage on Wikipedia content and editing patterns. Our research begins from Wiki Education's Wikipedia Student Program, and the participating courses and editors tracked by Wiki Education Dashboard. Higher education students have been participating in this program and making substantial English Wikipedia contributions since before the widespread adoption of LLM-based AI tools, and the adoption of AI tools for Wikipedia writing assignments since the launch of ChatGPT represents a natural experiment on the effects of AI usage on Wikipedia content.
As of 2026, we're working on building better ways of measuring fine-grained aspects of article quality — especially the dimensions of quality hypothesized to be impacted by AI usage — so that we can develop a clearer understanding of how AI is impacting the integrity of Wikipedia.
Methods
[edit]- Statistical analysis of Wikipedia contributions
- Analysis of Wikipedia contributions via AI detection services
- Article quality annotation by experienced Wikipedia editors
Timeline
[edit]- July 2025 — data collection from public Wiki Education course information and public Wikipedia edit history
- November 2025 — data collection via Pangram AI detector
- April 2025 — article quality annotation via https://wikiannotate.org/
Policy, Ethics and Human Subjects Research
[edit]- Princeton University IRB #19222, approved January 2026
Results
[edit]- The Impact of LLM Adoption on Student Writing: Lessons from Wikipedia Assignments - extended abstract from Wiki Workshop 2026
Resources
[edit]- Generative AI and Wikipedia editing: What we learned in 2025 - Wiki Education blog, January 29, 2026, by LiAnna Davis