AI-Based Profiling by Tax Authorities: Exploring GDPR Constraints and Explainability

Abstract

The article highlights growing legal concerns around the use of AI-based profiling by tax authorities, particularly regarding transparency and citizen rights. It emphasizes the General Data Protection Regulation’s (GDPR’s) prohibition of automated decision-making without human intervention, unless properly authorized and safeguarded. The CJEU clarified that ‘meaningful information about the logic involved’ must be provided in an accessible and intelligible way, not as complex formulas. Counterfactual explanations – offering ‘what if’ scenarios – are proposed as a promising technique to enhance transparency and meet GDPR requirements. The author advocates for more multidisciplinary collaboration to align AI-driven tax systems with legal principles, human rights, and technological advancement