Analysis of Hot Topics in AI Empowered Educational Technology Based on BTM Model

Against the backdrop of rapid AI advancements, AI-driven educational technology has emerged as a core driver of education's digital transformation, yet challenges such as short-text data sparsity and inefficient topic mining impede accurate analysis of educational hotspots. This study addresses this by analyzing 760 articles from China's CNKI and 120 international articles (largely from ICAIE) using the BTM (Biterm Topic Model), determining an optimal 8 topics (T=8) through perplexity calculations and visualization to explore deep semantic associations via word-pair co-occurrence in short texts. Results distill four key research directions: AI-driven educational digitization; ethical challenges and governance in intelligent education; personalized learning models and teaching innovation; and teacher professional development alongside educational ecosystem reconstruction. Through topic clustering and content analysis, the research reveals the collaborative evolution between technology-empowered education and its foundational educational principles, offering data-driven decision support for research and practice. Its innovation lies in leveraging the BTM model to overcome traditional topic models' limitations in short-text processing, providing a methodological framework for dynamic trend analysis in AI educational technology.