The proliferation of readily available Artificial Intelligence (AI) models and the recent promulgation of the EU’s Artificial Intelligence Act (AI Act) represent a paradigm shift that challenges established legal frameworks and societal norms. A defining feature of this shift is the AI Act’s intrusion into distinct 'regulatory spaces' historically governed by their own specialized rules and institutional actors. By extending its rights-based obligations into these domains, the AI Act inevitably generates friction with existing legal regimes. This normative conflict is poised to spur the development of new legal principles and demand novel policy responses. Established fields of law now face the imperative of reconciling their foundational doctrines with the Act’s mandates.

Among the domains affected is the ostensibly settled field of Value Added Tax (VAT). A particularly complex juxtaposition emerges between the EU VAT system and the AI Act. The Act challenges the EU VAT system on two levels. First, it designates certain AI practices as 'prohibited', rendering their supply and use unlawful ab initio because of the unacceptable risks they pose to fundamental rights. Second, it addresses the unique challenges presented by General Purpose AI (GPAI) models. While the provision of such models remains generally lawful, they function as core infrastructures for a vast array of downstream applications, some of which may fall within the prohibited category. GPAI models are further characterized as posing a 'systemic risk', an inherent quality rooted in their capacity to adversely affect democratic processes or public security on a large scale. This risk, by its very nature, cannot be entirely eliminated, but only mitigated through ongoing governance and control. The Act’s systemic risk framework thus aims not at eliminating risk altogether, but at safeguarding the Union’s fundamental values against large-scale societal harm.

This seminar contends that the AI Act introduces a novel legal paradigm predicated on the categories of unacceptable risk and systemic risk. This paradigm stands in fundamental tension with the EU’s VAT system and its predominantly economic rationale. Accordingly, the seminar discusses whether the principle of fiscal neutrality can legitimately compel the taxation of supplies that the Union itself has rendered unlawful and, more specifically, if this principle extends to AI practices prohibited as antithetical to the Union’s constitutional order, enshrined in Article 2 TEU and the Charter of Fundamental Rights, or to transactions inherently tainted by the materialization of systemic risk.

Bio:
Cristina Trenta is Professor of Public Law at Linnaeus University, Sweden. She also serves as Visiting Professor of Public Law at the School of Information Technology, Halmstad University, Sweden, where her work focuses on the legal implications of artificial intelligence, and is an affiliate researcher at the AI-Econ Lab, Örebro University. She holds a PhD in European Tax Law from the University of Bologna, Italy, and a second PhD in Commercial Law from Jönköping International Business School, Sweden. She is the author of articles and books on VAT, emergent technologies, and human rights, and has extensive long-standing EU-level experience as a two-time member of the European Commission's VAT Expert Group (2016–2019; reappointed 2022), as a member of the EU Observatory on the Online Platform Economy (2021), an advisory body to the European Commission that addresses issues related to the online platform economy, and as part of the expert group on next-generation digital technologies for the European Commission’s Directorate-General for Communications Networks, Content and Technology.

Registration

Please inform Liane Colonna (liane.colonna@juridicum.su.se) of your participation by the 20th of October.

Warm welcome!

Liane Colonna and Stanley Greenstein