Intelligenza artificiale e costituzionalismo: il modello europeo

Artificial intelligence constitutes a challenge to traditional legal categories, including from the perspective of European constitutionalism. Within this framework, the European Union has progressively developed a regulatory model which, although not exempt from rigidities and still marked by margins for simplification, remains anchored in the centrality of the person and in the protection of rights. Moving from the broader European corpus iuris of the digital sphere to the AI Act, the article reconstructs a regulatory trajectory in which the risk-based approach acquires the meaning of a form of “constitutionalism applied to the digital domain”. The European model is then compared with the prevailing approach in the United States, characterised by absent or fragmented regulation, market-oriented logic, and the safeguarding of veritable private digital quasi-states. In this respect, the specificity of the European model lies not only in its regulatory approach, but also in its attempt to subject technological powers to a public-law grammar shaped by controls and guarantees. Despite the difficulties and the pressures arising from global competition, the European model emerges as a possible ordering criterion for reconciling innovation, fundamental rights, and democratic guarantees, as well as a basis for the elaboration of shared minimum principles within the emerging global governance of artificial intelligence.