"Gamifying XAI: Enhancing AI Explainability for Non-technical Users through LLM-Powered Narrative Gamifications" is a project that explores how gamification and narrative-driven interactions, powered by Large Language Models (LLMs), can enhance AI explainability for non-technical users. The study focuses on integrating LLMs into Explainable AI (XAI) visualization technique, aiming to improve XAI visualizations' meaningfulness and relatedness for non-technical users. Our prototype introduces LLM-driven conversational NPCs that guide users through complex AI concepts and XAI visual encodings, helping them understand things like model prediction process and decision boundaries in a more intuitive way.
- Links:
- Video Figure: Youtube Link
- Video Demo: Youtube Link
- Recognition & Outreach
- Accepted by CHI'25 as a LBW paper!
- Core Features
- Introduces gamification elements into XAI visualizations, specifically focusing on improving non-technical user understanding.
- Integrates Large Language Models (LLMs) to create narrative-driven NPCs that explain AI models and visualizations.
- Explores interactive t-SNE projections to allow users to explore model embeddings and understand AI decision-making processes.
- Produces design implications of LLM-driven gamification in improving explainability and reducing cognitive load for non-technical AI users.
- SkillsPython, t-SNE, XAI Visualization, Large Language Models (LLMs), Gamification, AI Explainability, OpenAI API, Javascript, D3.js
- AuthorsYuzhe You, Jian Zhao
- Paper LinkGamifying XAI: Enhancing AI Explainability for Non-technical Users through LLM-Powered Narrative Gamifications
- KeywordsHuman-Computer Interaction, Information Visualization, Explainable AI, Gamification, t-SNE, Large Language Models, XAI Visualization