A BIBLIOMETRIC ANALYSIS OF AI-DRIVEN PROCEDURAL CONTENT GENERATION IN VIDEO GAMES: KEY CONTRIBUTORS, THEMATIC TRENDS, AND COLLABORATION NETWORKS FROM 2008-2025

Abstract

AI-driven Procedural Content Generation via Machine Learning (PCGML) is profoundly impacting the video game industry, yet its academic literature remains highly fragmented. This creates a significant challenge for researchers and practitioners needing to track trends, identify foundational work, and find expert collaborators. This study attempts to address this gap by providing a data-driven map of the PCGML research field through a rigorous bibliometric analysis. Following PRISMA guidelines, this study analyzed 2,184 curated documents from the Scopus database (2008-2025). A quantitative analysis using the bibliometrix R package examined the field's conceptual, intellectual, and social structures. Revealing a clear evolution from classic AI toward a modern core dominated by deep learning and reinforcement learning. The United Kingdom was identified as the leader in research impact, while network analysis uncovered a core group of highly influential authors. This paper provides a valuable roadmap for academics by highlighting research gaps and offers industry practitioners strategic insights for technology adoption and collaboration, effectively bridging the gap between academic theory and practical application.