Educational AI Chatbots and Risks to Teaching and Learning: A Systematic Review

• In Event: AI Literacy in Education: Exploring Pedagogical Approaches, Ethical Considerations, and Diverse Student Perspectives

Abstract

With generative AI posing new opportunities and challenges to education, this study investigates how risk has been conceptualized and studied in educational chatbot research. A systematic review of peer-reviewed articles was conducted. The results: (1) identify five categories of risks: pedagogical, technological, ethical/social/cultural, political, and legal; (2) reveal contextual differences in risk emphasis, such as cultural and legal risks in regions and interface design concerns in special education; (3) highlight the unbalanced empirical research across educational levels and a lack of medium-to-large sample studies; (4) synthesize the mitigation strategies from educational intervention, technical advancement in related computer programs, and the creation of governance rules. Suggestions for further research include exploring underrepresented contexts, conducting larger-scale empirical studies, and optimizing mitigation strategies.

Authors

• Yaolan Weng, The Ohio State University

• Rick J. Voithofer, The Ohio State University