Mapping the Landscape: A Bibliometric Analysis of Rating Agencies in the Era of Artificial Intelligence and Machine Learning

Abstract In the ever-evolving financial landscape, the integration of artificial intelligence (AI) and machine learning (ML) is revolutionising creditworthiness assessment. The vast body of literature on credit rating indicates a growing prevalence of these techniques in the rating processes. Although these methods boast high predictive accuracy, concerns about their robustness, equity, and explainability affect the confidence of various parties in rating agencies. This comprehensive study explores the dynamic intersection of these cutting-edge technologies with rating agencies, presenting an in-depth literature review employing a bibliometric analysis that uses the Bibliometrix and Biblioshiny packages from R. The paper makes a significant contribution by analysing the literature across three prominent databases: Web of Science, Scopus, and arXiv. The empirical findings indicate that despite a recent growing interest, the relatively limited number of documents implies that, while there is a wide literature about credit rating in general, when it comes to rating agencies, the literature is much more limited. This limitation may stem from a certain lack of transparency in the methods and processes used by rating agencies and the complex nature of these entities. The literature witnessed growth after the 2008 global financial crisis, where rating agencies faced significant criticism, and post-pandemic, indicating a need for more adaptable and precise ratings. The examination of the topic reveals a recent shift in focus within AI-driven rating agencies towards accountable governance. While traditional attention persists on artificial intelligence techniques and finance, the emerging emphasis on ethical considerations, societal impacts, and performance evaluation underscores a changing landscape. This transition underscores the growing importance of integrating ethical considerations and societal impacts into the operational frameworks of AI-powered rating agencies, emphasising the necessity for responsible and transparent decision-making practices.