GenAI: a scientometric analysis of research trends using biblioshiny and VOSviewer

Abstract

Generative Artificial Intelligence (GenAI) has rapidly emerged as a transformative technology shaping diverse domains such as education, defense, healthcare, industry, tourism, and knowledge management. Despite its expanding influence, the scholarly landscape of GenAI research remains fragmented, making it challenging to trace its growth patterns, collaboration networks, and thematic evolution. To address this gap, this study conducts a comprehensive scientometric analysis to systematically map the intellectual structure and research dynamics of the field. The objective is to identify the most influential authors, institutions, countries, and thematic clusters, thereby providing a holistic view of the development of GenAI research. The study analyzes 9612 documents retrieved from Scopus and Web of Science databases (1986–2025, based on data availability). Data screening followed the PRISMA protocol to ensure transparency and rigor. Advanced bibliometric tools viz. Biblioshiny (R-based Bibliometrix) and VOSviewer were employed to examine productivity indicators, visualize co-authorship networks, and uncover thematic structures using keyword co-occurrence, trend analysis, and Latent Dirichlet Allocation (LDA) topic modeling. The results reveal a steep escalation in GenAI-related publications, increasing by 3164% in Period 2 (2022–24) compared to Period 1 (2012–22), reflecting the disruptive rise of text-generation tools such as OpenAI’s GPT series, ChatGPT, and Gemini. The United States, China, India, and the United Kingdom dominate research output, with Harvard University, Nanyang Technological University, and the University of California leading institutionally. LDA topic modeling identifies ten major thematic clusters, spanning technical design, image synthesis, and large language models to applications in education, healthcare, legal governance, and ethical inquiry. Topic prevalence highlights governance, ethics, and risk (20%) along with AI-driven content creation (17%) as the most explored themes, followed by educational applications (14%). Lotka’s Law analysis further reflects the characteristics of an emerging domain, with numerous occasional contributors and a smaller, prolific core group driving the research front. This study represents a comprehensive scientometric investigations of GenAI to date, integrating performance metrics, network visualizations, and advanced topic modeling. By mapping the intellectual structure, global collaborations, and evolving research trajectories, it offers actionable insights for scholars, practitioners, and policymakers seeking to navigate and contribute to this rapidly advancing field. The analysis reveals two dominant directions in GenAI research: governance across sectors and technical or application-specific innovation.

Ähnliche Arbeiten