Designing XAI Chatbots to Enhance Learner Self-efficacy in Education
摘要
Self-efficacy is crucial for self-regulated learning (SRL), helping learners overcome obstacles and achieve educational goals. This study introduces ALLURE, a multimodal AI platform with an XAI-driven chatbot for teaching Rubik’s Cube solving to enhance self-efficacy. We analyzed interactions of college students using interviews, think-aloud protocols, and observational notes. Preliminary findings show that while XAI-driven chatbots can boost self-efficacy, their effectiveness varies. These results highlight the need for educational chatbots to cater to diverse learning needs. The study provides insights into developing XAI-driven chatbots with transparent, personalized interactions to support self-efficacy and competence in educational settings.