Artificial intelligence in education: mapping research trends, engagement and cognitive disengagement with evidence from global and Asian contexts

Purpose This study aims to examine the intellectual structure, evolution and emerging trends of artificial intelligence (AI) in education, with a specific focus on student engagement and cognitive disengagement, within a rapidly expanding and technology-driven research landscape. Design/methodology/approach A bibliometric analysis was conducted using the Scopus database, covering journal articles published between 2015 and 2026. A total of 3,861 publications were analyzed using performance analysis and science mapping techniques. Tools such as Bibliometrix and R software were employed to examine publication trends, citation patterns, keyword co-occurrence, thematic structures, and disengagement-related research patterns. Findings The results reveal significant growth in AI-in-education research, particularly after 2023, driven by the emergence of generative AI technologies such as ChatGPT. Thematic analysis indicates a shift from technology-oriented research toward learner-centered constructs, with student engagement acting as a key bridging concept. It is important to distinguish two related but separate claims; while a substantial proportion of the corpus (61–73% annually) addresses constructs adjacent to disengagement such as motivation, attention, and engagement cognitive disengagement as an independent, explicitly labelled construct receives comparatively limited direct attention (2.2% of the disengagement sub-corpus). This distinction reconciles the apparent tension between the high proportion of disengagement-adjacent literature and the claim that disengagement remains underexplored as a standalone variable. The findings also highlight a geographically concentrated research landscape, with strong contributions from Asian countries, particularly China and the Asia-Pacific region. Research limitations/implications The study is limited to Scopus-indexed journal articles and English-language publications. Future research should explore cognitive disengagement through empirical and longitudinal studies, particularly in diverse educational and regional contexts, including underrepresented areas within Asia. Practical implications The findings emphasize the need for educators and policymakers to adopt a balanced approach to AI integration, ensuring that technological advancement does not compromise critical thinking, attention, and deep learning. Originality/value This study contributes to the literature by integrating bibliometric mapping with a disengagement-focused perspective, offering a novel understanding of how AI technologies influence both engagement and cognitive disengagement in education.