Mapping the Intellectual Structure of Artificial Intelligence for Tax Compliance Enhancement: A Bibliometric Review

Abstract

Background: Artificial Intelligence (AI) has revolutionized tax compliance, improving fraud detection, risk assessment, and voluntary compliance. The knowledge landscape, intellectual bases and the trends in this nearing field, however, are not sufficiently well articulated and require a systematic mapping. Methods: The structured search strategy was applied to the PRISMA methodology to retrieve literature on AI and tax compliance from Scopus. The number of records analyzed is 527 records published between 2004 and 2026, and it was done through Biblioshiny and VOSviewer. The publication trends, prolific authors, journals, countries, and citations, were analyzed to explore the performance, and science mapping included co-citation, bibliographic coupling, keyword co-occurrence, and thematic evolution. Results: Research has been growing at an exponential rate since 2020, thanks to developments in machine learning, big data analytics and digital tax administration. Leading contributors became China, Germany and the United States. Focusing on the dominant themes in the field of taxation, artificial intelligence, machine learning, fraud detection and data mining. The growing research shows the importance of explainable AI, governance, transparency, public trust and ethical AI adoption. Conclusion: The research suggests a five-step framework for implementing AI for tax compliance: risk concept formulation, data integration, model selection, governance and explainability, and continuous monitoring. Novelty: This review article not only offers a comprehensive bibliometric mapping of AI and tax compliance, but also combines performance mapping with science mapping, and selects emerging themes and new research directions.