KnowTI: Bridging the Knowledge Gap in Information Communication through Interactive Visual LLM Summarizations

Abstract

The gap between expert knowledge and layperson understanding often poses challenges to the effective communication of unfamiliar concepts. To address this issue, we introduce a semi‐automated, interactive visual summarization framework with modularity and user control, KnowTI, which leverages Large Language Models (LLMs) and visualization methods to facilitate knowledge transfer. Our approach aims to narrow the knowledge gap of laypersons in comprehending complex information as presented by experts, while remaining engaged, making our solution applicable to a variety of information communication contexts—including, but not limited to, educational videos. We validate the effectiveness of KnowTI through a comparative study with 40 participants, contrasting our approach with traditional video‐based learning. Findings indicate that KnowTI not only enhances knowledge acquisition and engagement, but also supports laypersons in understanding complex, domain‐specific topics, bridging the knowledge gap between experts and non‐experts. All supplementary material, including the code, analysis, and evaluation, is available on GitHub .