Transforming Lecture Videos into LLM-Enhanced Immersive Learning Experiences

Video lectures have become a dominant medium for distance learning due to their accessibility and flexibility. However, their largely passive format limits opportunities for interaction, making it difficult to sustain learner engagement and support active learning. To address this limitation, we describe a workflow for transforming video lectures into interactive and immersive learning experiences. Following this workflow, a lecture setting is replicated from an input video and augmented with a pedagogical agent driven by a prompt-engineered large language model (LLM). The lecture experience includes a Q&A mechanism that enables students to ask questions and receive responses from the agent. We evaluated the immersive learning experience (Immersive) developed with our workflow in a study against a video lecture (Video) and a desktop-based interactive version (Desktop). Our results show that the interactive conditions (i.e., Desktop and Immersive) significantly increased engagement compared with the passive condition (i.e., Video), while only the Immersive condition increased intrinsic interest. However, both interactive conditions also increased reported mental effort and perceived pressure relative to the passive condition. Importantly, participants in the Immersive condition achieved significantly higher learning gains than those in the Video condition, whereas no significant differences emerged between the Desktop and the other conditions. These findings highlight both the potential benefits and the practical trade-offs of transforming passive instructional media into interactive and immersive learning experiences. Based on these insights, we derive design guidelines for converting conventional video lectures into interactive and immersive learning experiences.