Mapping the multidimensional trend of generative AI: A bibliometric analysis and qualitative thematic review

Abstract

Generative artificial intelligence (AI) represents an increasingly popular topic that is visible even in most research areas within the social sciences and humanities fields. However, little attention has been paid to the knowledge dimensions reflecting the potential macro-social implications of generative technologies. This study utilizes a two-fold methodology, consisting of a bibliometric analysis of articles published in the last decade (N = 484) and a subsequent qualitative thematic review of the most influential articles in each research area (N = 246). The objective is to investigate the main conceptual dimensions associated with generative AI in the social sciences. Applying a thematic analysis framework, we notice that the most popular dimensions are technological, ethical, and social. These dimensions primarily focus on investigating the implications of the generative use of AI on employees in professional sectors as well as on students and teachers in the educational environment. Moreover, the political dimension reflects macro-social consequences on governance and legal components related to ensuring social protection for professions that risk becoming obsolete due to the widespread adoption of ChatGPT-type technologies. Overall, our research emphasizes concrete scholarly tensions through which generative AI-based technologies are predominantly encouraged in the educational and organizational sectors, but the potential risks associated with copyright infringement and job loss might constitute important drivers of social change. We also notice that a Foucauldian power/knowledge framework would prove useful in understanding the underdiscussed effects of generative AI on the societal/macro level.