Artificial Intelligence in Tax Compliance and Evasion Mitigation: Trends, Mechanisms, and Institutional Implications

Abstract

This study presents a systematic literature review of 68 peer-reviewed articles (2015–2025) on artificial intelligence in tax compliance and evasion mitigation. Using the PRISMA 2020 protocol and textometric analysis via IRAMUTEQ software, we map publication trends, geographic distribution, and three research paradigms: machine learning and predictive modeling; artificial intelligence, technology and tax compliance; and government, financial development and revenue administration. The CIMO (Context–Intervention–Mechanism–Outcome) framework structures our synthesis of how institutional conditions shape intervention design and why identical technologies produce divergent outcomes across settings. While existing reviews have focused primarily on detection metrics without theorizing institutional boundary conditions, behavioral dynamics without addressing governance capacity, or ethical deficits without a theoretical framework, this study constructs the Adaptive AI Tax Compliance Framework (AAITCF), a context-sensitive implementation roadmap differentiated across three institutional maturity tiers. The results indicate that AI achieves high detection accuracies in digitally mature economies, yet effectiveness is contingent on data quality, governance capacity, and organizational readiness. Developing countries face structural asymmetries, infrastructural deficits, and human capital gaps that constrain algorithmic performance even where technical sophistication is high. The AAITCF treats context as constitutive of intervention effectiveness and identifies underexplored areas regarding causal pathways from AI deployment to long-term institutional change, taxpayer trust, and equitable fiscal governance.