Exploring How Agent Voice Accents Shape Human-AI Collaboration in K-12 Group Learning

A study published on arXiv explores the impact of voice accents in GenAI conversational agents on K-12 group learning dynamics. The research involved a mixed-methods approach with 33 teachers, revealing that accents influenced participants' perceptions and roles of the agents; specifically, a British-accented agent was viewed as a tool, while Indian and African American accents fostered more peer-like interactions. These findings highlight the importance of sociolinguistic design in AI agents for enhancing collaboration and trust in educational settings, suggesting that culturally inclusive design can improve group learning outcomes.