Exploring Retrieval-Augmented Generation (RAG)-driven wiki edit preparation: early insights, challenges, and potential
This session will explore how RAG was integrated into wiki edit preparation, highlighting methodological choices, tools used, and key challenges encountered in an educational setting.
Abstract
This session presents early experiments using a Retrieval-Augmented Generation (RAG) approach to prepare wiki edits within an open knowledge ecosystem. These experiments are part of a larger project on AI-assisted collaborative editing and represent the initial stages of the work outlined in this publication. This pilot phase was conducted with pedagogy students during the first semester of 2025. The session will explore how RAG was integrated into wiki edit preparation, highlighting methodological choices, tools used, and key challenges encountered in an educational setting. Preliminary findings suggest that RAG can enhance content preparation by improving the review of academic sources, breaking language barriers, and enabling a multimedia approach. These early insights contribute to the broader conversation on AI-enhanced open knowledge practices, shaping future developments in post digital editing environments and informing the ethical, technical, and pedagogical dimensions of AI-assisted knowledge production in education.