AI in Education: Review and Prospect

Abstract

The application of artificial intelligence in education is undergoing a paradigm shift from ”tool embedding” to ”system reshaping”. This study uses 1,181 relevant articles from the 2013-2025 Web of Science core collection as samples, employing a combination of bibliometric analysis and thematic content analysis to map the spatiotemporal evolution, knowledge networks, and research hotspots in this field. The findings reveal that annual publications surged from 1 in 2013 to 715 in 2025, with citation counts reaching 8,385. Since 2024, the field has demonstrated ”dual leaps in quantity and quality”. Major research contributors include China, USA, and Australia, with papers predominantly published in the journal Education and Information Technologies. Research themes can be categorized into six major sectors, extending to eight application scenarios such as language learning, basic education, special education, and teacher development. Current studies predominantly adopt cross-sectional designs, lacking long-term tracking and causal inference, while insufficient attention is given to rural areas, low-resource languages, and special education contexts. Ethical risk governance frameworks also require further refinement. This paper proposes future research directions focusing on three dimensions: paradigm innovation, contextual expansion, and ethical governance, aiming to provide references for advancing AI in Education research and informing policy practices.