The Evolution of Generative AI in Science Educational Contexts: A Bibliometric Study
Risya Pramana Situmorang (a*)- Panggih Priyambodo (b)

a) Biology Education Study Program, Faculty of Biology, Satya Wacana Christian University, Salatiga, Indonesia.
*risya.situmorang[at]uksw.edu
b) Biology Education Study Program, Faculty of Teacher Training and Education, Universitas Lampung, Lampung, Indonesia

Abstract

The rapid integration of Generative AI (GenAI) into science education presents a paradoxical condition- on one hand, the technology offers transformative potential for creating adaptive and personalized learning environments, yet on the other hand, it also introduces complex challenges such as academic dishonesty, algorithmic bias, and over-reliance, which potentially undermine the primary goals of science education: conceptual understanding and critical thinking. This study aims to provide the first comprehensive mapping of the evolution of GenAI research in science education over two decades (2004-2025) to identify global publication trends, research collaborations, hotspot topics, and thematic developments. Utilizing Bibliometrix in RStudio and VOSviewer, this research analyzed 7,489 relevant articles from Scopus data. The results reveal a significant surge in publications since 2018, identify dominant research clusters (e.g., large language models, ethical issues), and map the leading contributing countries and institutions. The use of GenAI is largely directed toward personalizing learning and enhancing student understanding, such as through digital textbooks and educational chatbots. The primary contribution of this research is an objective foundational framework that helps prevent future fragmentation of studies, supports informed policymaking for science education in the AI era, and establishes a clear agenda for subsequent research in this rapidly evolving field.

Keywords: Generative AI- Science Education- Bibliometric Analysis

Topic: Science Education