Artificial Intelligence (AI) has increasingly been integrated into educational systems to enhance student performance analysis, early warning systems, and academic risk prediction. Recent advancements in machine learning, data analytics, and educational data mining have enabled institutions to identify at-risk students and provide timely interventions. While existing studies have primarily focused on predictive models and algorithmic accuracy, limited attention has been given to understanding the overall research structure, thematic evolution, and intellectual development of AI-driven academic risk assessment. The existing literature remains fragmented, with a lack of comprehensive analysis of research trends, influential contributors, and emerging themes in this domain. Furthermore, there are relatively few bibliometric studies that systematically examine the intersection of artificial intelligence, academic performance, psychological stress, and family-related factors affecting student outcomes. This study aims to conduct a comprehensive bibliometric analysis of research on AI-driven academic risk and student performance. Using data extracted from major academic databases, the study examines publication trends, key authors, influential journals, collaboration patterns, and thematic developments within the field. The findings reveal a rapid increase in research output, particularly after 2022, driven by advancements in machine learning and educational analytics. Additionally, the study identifies key research gaps, including limited integration of psychological and family-related variables, ethical concerns in AI-based decision-making, and insufficient interdisciplinary collaboration. This research contributes to the literature by providing a structured overview of the evolution of AI-driven academic analytics and highlights future research directions. The insights will assist researchers, educators, and policymakers in developing more effective and ethical AI-based educational interventions.
J et al. (Fri,) studied this question.