“Error 404 – Match not found”: tax enforcement and law enforcement in the EU Artificial Intelligence Act

• Autores: David Hadwick
• Localización: Eucrim: the European Criminal Law Associations' fórum, ISSN 1862-6947, Nº. 1, 2023 (Ejemplar dedicado a: Artificial intelligence), págs. 55-60
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Dialnet Métricas: 1 Cita
• Resumen

  • In EU Member States, tax administrations are the public organs that make most use of artificial intelligence (AI) and machine learning (ML) systems to perform State prerogatives. At least 18 EU Member States frequently use AI tax enforcement systems. In certain areas of taxation, such as value-added tax, AI and ML are already used throughout the EU. These systems perform a relatively broad range of tasks, reflecting the wide array of prerogatives of the administration itself. Generally, these different systems can be categorized into two archetypes: coercive and non-coercive AI systems. While non-coercive AI tax systems do not generate significant risks of conflict with taxpayers’ fundamental rights, coercive AI tax systems used for tax enforcement bring about serious risks of conflict with taxpayers’ fundamental rights and tax procedure as a whole. These risks have already materialised in a number of cases and have even led to serious scandals, such as RoboDebt and the toeslagenaffaire.

    Yet, substantial confusion exists around the treatment of AI tax enforcement systems in the upcoming Regulation laying down harmonised rules on artificial intelligence (‘EU AI Act’) and whether these systems will be qualified as high-risk. Recital 38 of the current draft prescribes that systems used by tax administrations specifically for administrative purposes should not be viewed as high-risk AI law enforcement systems. While prima facie logical, distinguishing between administrative and law enforcement purposes is bound to be an impractical and arbitrary exercise. Law enforcement is becoming increasingly integrated through the involvement of administrative authorities and private actors, precisely because of the use of AI. In such contexts, the boundaries between administrative and penal processes are blurred and will generate confusion. By remaining attached to that anachronistic distinction, Recital 38 not only replicates that confusion but will exacerbate its effects.