‘Behind the One-Way Mirror : Reviewing the Legality of EU Tax Algorithmic Governance’

Tax algorithmic governance has surged from a handful of Member States a decade ago, to a majority of tax administrations in the European Union (EU) integrating artificial intelligence (AI) systems. In light of the ever-increasing volume of tax returns and tax documentation to be processed, the digital transformation of the administration has become an imperative. Yet, cases such as systeem risico indicatie (SyRI), the toeslagenaffaire and eKasa show that automation poses risks to taxpayers. These cases cast doubts on the secretive nature of the tax administration’s prerogatives and the information and communications technology (ICT) tools used to perform its missions, begging the question of what lies behind the administrations’ one-way mirror. Section 2 presents the current state of use of AI tax systems on the basis of a synthesized literature review of publicly available data, mapping how many and what EU States have integrated AI systems. Section 3 examines compliance of the AI systems identified with the principle of legality. This research finds that while EU Member States have heavily invested into the integration of AI tax systems, very few have adopted specific norms to mitigate the risks to taxpayers’ fundamental rights. As a result, tax algorithmic governance is creating a noticeable gap in the protection of taxpayers’ rights