Research:The Impact of WikiVault on Korean Wikipedia
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.
WikiVault is a GenAI-assisted Wikipedia editing tool that has been deployed on the Korean Wikipedia since May 2025.
It was developted by Korean Wikipedian Ted Yoo, one of the project's collaborators, with the support from the Korean WikiMedia Association.
The tool currently provides two main functions:
- Translation – translating English articles into Korean
- Article drafting – generating initial article drafts from user prompts
More details about the tool's functionality can be found on this page.
Since its deployment, there have been active community discussions within the Korean Wikipedia about the use of WikiVault - ranging from policy considerations to values and concerns around AI-assisted editing. Across many language editions of Wikipedia, communities are also exploring how to thoughtfully integrate GenAI tools into their workflows. The Korean experience with WikiVault can provide a valuable opportunity to learn from a real-world deployment and contribute data-driven, community-centered insights to these broader conversations.
Therefore, we aim to systematically understand the impact of this tool on the Korean Wikipedia - combining community perspectives with analysis of editing data and behavioral patterns. By doing this, the project aims to:
- Support the Korean Wikipedia community in shaping policies and improving the design of WikiVault in line with community values.
- Share lessons learned and insights with other Wikipedia communities exploring the adoption of GenAI-assisted editing tools.
Methods
[edit]This project uses a mixed-methods approach that combines:
- Quantitative analysis of edit histories, such as tags (e.g.,
wikivault,mw-rollback,mw-reverted), and user activity before and after WikiVault deployment. - Qualitative analysis of editors’ experience and perceptions through semi-structured interviews.
Timeline
[edit]We are currently actively working on this project, currently in data analysis phase (quantitative), and interviewee recruitment phase (qualitative).
Policy, Ethics and Human Subjects Research
[edit]The project includes both qualitative and quantitative components:
- Quantitative analysis:
This part of the study relies on Wikimedia data dumps. As such, it involves no direct interaction with editors and does not interfere with ongoing Wikipedian work.
- Qualitative study (semi-structured interviews):
This component has been reviewed and approved by the Princeton University Institutional Review Board (IRB project #18389, approved on 16 October 2025). Participation in the interviews will be entirely voluntary, and informed consent will be obtained from all participants prior to data collection. Data will be stored securely, and results will be reported in aggregate to protect individual confidentiality.
Together, these approaches ensure that the study complies with ethical research standards and minimizes disruptions to the Wikimedia community.
Results
[edit]Results and analyses are currently in progress. Preliminary findings will be shared on this page once available.
Resources
[edit]We plan to share replication materials (code, notebooks, and documentation) as the project progresses to ensure transparency and reproducibility.