ChatGPT and AI Chatbots in Education: An Umbrella Review of Systematic Reviews, Scoping Reviews, and Meta-Analyses
Milakis Emmanouil, Constantina Corazon Argyrakou, Alexandros Melidis, John Vrettaros
Abstract
This umbrella review synthesizes findings from 41 systematic reviews, scoping reviews, and meta-analyses on the use of ChatGPT and similar large language model (LLM)-based chatbots in education. It provides a critical analysis of their pedagogical applications, benefits, limitations, and associated ethical and policy issues across diverse educational levels and domains. The evidence reveals that chatbots are predominantly implemented in higher education, with growing use in medical, STEM, and language learning contexts. Reported benefits include personalized learning, enhanced writing and critical thinking skills, and increased learner autonomy. However, significant concerns persist regarding the reliability of AI-generated content, overreliance by students, academic integrity, and institutional preparedness. The review highlights methodological gaps in current research, such as a lack of longitudinal studies and limited attention to underrepresented populations and educational settings. The findings aim to inform evidence-based decision-making for educators, researchers, and policymakers navigating the integration of AI chatbots into formal education systems.