Artificial intelligence in tax compliance: A bibliometric mapping and research agenda (2015–2025)

Abstract

While AI is integral to the tax compliance ecosystem, academic research is scattered across tax enforcement, governance, and sustainability. This bibliometric analysis of 1,803 Scopus-indexed articles from 2015 to 2025 shows that the field is not yet mature, but is undergoing rapid thematic change, with three waves, enforcement, governance, and sustainability, compressed into a decade. The central efficacy-legitimacy challenge is clear: AI improves anomaly detection and audit targeting, but algorithmic opacity can undermine taxpayer trust. The geographic inversion is also telling: Ukraine, Jordan, and Malaysia have more studies than OECD operational centers, suggesting reform urgency rather than administrative maturity. The study introduces the ATCGM, combining effectiveness, legitimacy, and sustainability, and identifies five empirically grounded research gaps. The practical implications are real: Peru’s e-invoicing rollout increased declared VAT liabilities by over 5% in its first year, but the governance conditions behind such gains remain under-researched.