Artificial Intelligence and Regulatory Enforcement: A Comparative Research Design of Tax Administrations in Argentina, Brazil, Estonia, and the Netherlands

Tax authorities across the world are adopting artificial intelligence systems to decide whom to audit, yet enforcement theory was built on assumptions about human interaction that these systems disrupt. This research design studies how AI reshapes the regulator-regulatee relationship through four tax administrations with different levels of algorithmic deployment and institutional oversight: ARCA in Argentina, Receita Federal in Brazil, the Estonian Tax and Customs Board, and the Dutch Belastingdienst. The central hypothesis is that AI's effect on enforcement posture is not uniform but mediated by institutional oversight. The design combines within-case process tracing with structured cross-case comparison, supported by primary sources. The work contributes a bridge between classic enforcement theory and algorithmic governance, a comparative approach with two Latin American and two European cases, and practical implications for sequencing AI adoption, identifying institutional pressure points, and designing evaluable reforms.