Harnessing Topic Modeling to Investigate the Intersection of Accounting and Artificial Intelligence through Systematic Literature Mapping

Abstract

Previous research has often suggested that various accounting functions could be replaced by Artificial Intelligence (AI) and related technologies. However, more recent studies increasingly recognize AI's potential to enhance value within accounting practices and organizations. Scholars and experts have called for more extensive research into the relationship between accounting and AI, emphasizing the importance of adopting a multidisciplinary approach in this field. This paper employs topic modeling, specifically Latent Dirichlet Allocation (LDA), to systematically analyze the existing literature on AI and associated technologies within accounting. By applying LDA to the abstracts of 930 peer-reviewed articles from diverse academic fields published between 1990 and 2023, the study identifies key themes and trends in the discourse around accounting and AI. The results indicate that previous literature reviews using conventional methods may have overlooked important aspects of this rapidly evolving area. The analysis reveals eleven distinct topic clusters that together form a detailed map of the current research landscape. These findings not only broaden understanding of accounting and AI scholarship but also offer a structured framework for guiding future investigations. Additionally, this research represents one of the pioneering uses of probabilistic topic modeling techniques within the accounting literature.