A Comprehensive Bibliometric Analysis of Artificial Intelligence Research in the Field of Science Education

Abstract

Today, the importance of artificial intelligence in science learning and teaching is rapidly increasing. The growing interest in this field and the resulting increase in academic publications on the subject make it challenging to understand its progress and trends on a global scale. Furthermore, a literature review reveals a notable lack of studies that offer a comprehensive perspective, reflecting the current state and research trends in this field. Therefore, this study aims to analyze the current state, evolution, and important research trends in studies on artificial intelligence in science education from 1985 to 2024, utilizing bibliometric methods. To this end, a total of 169 articles were analyzed from the Web of Science database using specific keywords. Analytical tools such as VOSviewer and SciMAT software were used for data visualization. The results indicate that research on artificial intelligence in science education from 1985 to 2024 has developed irregularly, with significant growth occurring in recent years. The country with the highest citation and production levels in this research field is the United States. The most productive journals in the area are the Journal of Science Education and Technology, Frontiers in Education, and the Journal of Research in Science Teaching. The leading authors are Cooper, G., and Zhai, X. Keyword analysis showed that “science education,” “computer science education,” “machine learning,” “artificial intelligence assessment,” “ChatGPT,” and “learning analytics” are among the most frequently used terms and highlight emerging thematic clusters. Furthermore, this analysis showed that while artificial intelligence research in science education was initially more limited and focused on technology-related themes, it has recently shifted toward a research direction that includes learning analytics, interactive learning environments, computational thinking, and large language models. The results offer a guiding framework and valuable insights for practitioners and education researchers seeking direction in the evolving landscape of artificial intelligence in science education.