Oak Story: Improving Learner Outcomes with LLM-Mediated Interactive Narratives
summary
A mobile app teaches 4th–6th graders about oak woodlands using a scene-graph narrative, AR activities, and a learning-goal-directed LLM architecture that chooses one goal at a time, calls discrete actions (like start AR or transition scenes), and uses retrieval-augmented generation over a manually curated knowledge base. A between-subjects study with N = 47 compares adaptive progression to a linear control and finds statistically significant increases in learning gains, engagement, and perceived agency.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:06Project overview
- 2:08Demo and app features
- 4:17LLM architecture details
- 7:05Experiment design
- 8:55Results and findings
- 10:20Conclusion and summary
- 10:40Q&A begins
related talks
01h 23mEmpowering Learning: IT, Education, and Ethical Considerations
41mBringing stories to life with AI, data streaming and generative agents
13mStrategies for Enhanced Personal Knowledge Management
52mAI-Powered Learning Pathways & Skills Development
01h 04mDesigning Future-Ready Classrooms: Leveraging AI & OER to Support Diverse Learners
13mInteractive Learning and User‑Centered AI: A Retrospective
26mGrappling with the Bitter Lesson
01h 28mStoopid Questions - about learning, teaching and AI