Unveiling Trends in Deepfake Research

A Global Bibliometric Perspective

Abstract

Deepfake technology leverages artificial intelligence to generate highly realistic, yet manipulated, videos that can potentially mislead or misinform the public, thereby creating ethical and social challenges. This study aims to analyse the evolution, productivity, and collaborative networks in deepfake research from 2018 to 2023, investigating the contributions of various authors, institutions, and countries to understand key trends and interdisciplinary impacts in this growing field. Using bibliometric methods, the research data was collected from Scopus database and analysed 1,548 publications across 700 sources. Tools such as RStudio and VOSviewer were employed to assess research performance and visualise co-authorship and citation networks. The analysis identifies computer science, engineering, and mathematics as leading fields, with China and the United States as primary contributors. Collaborative dynamics reveal substantial partnerships among these nations and with other active contributors like India and Singapore. Major research clusters were observed around influential journals, such as Lecture Notes in Computer Science. The study also provides an in-depth examination of citation patterns, revealing a high concentration of citations in specific journals and among leading authors. Furthermore, co-authorship networks illustrate the growing international collaborations in the field, with increasing interdisciplinary cooperation. The growth rate of deepfake research has surged, reflecting the rising importance of this topic in various domains. This study provides a comprehensive overview of the academic landscape of deepfake research, emphasising the thematic and collaborative structure. Through network visualisation and thematic analysis, the study underscores deepfake research’s global and interdisciplinary nature, setting a foundation for future explorations.