Algoritmi e diritti. Trasparenza, non-discriminazione e proprietà degli output dell’intelligenza artificiale

Daniele Marongiu

2025-01-01

---

Abstract

Artificial Intelligence has entered the lives of individuals and the dynamics of human relationships. After nearly a century of theoretical development, the conditions have been created for its entry into our daily lives. Until the last decade, the commonly used algorithms represented processes that faithfully reproduced completely predetermined human instructions. Today, through machine learning systems, computers are capable of. learning, detaching themselves from the programmer’s instructions, and acting autonomously. This book addresses how the law is addressing this change by creating a system of rules that define rights, duties, and powers in relationships mediated by algorithms, both traditional and based on machine learning. These pages focus on three areas of analysis. The first concerns transparency, and therefore the development of rules that prevent asymmetric knowledge about the content of algorithms from generating social imbalances. The second concerns non-discrimination, thus the legal community’s attention to the risk that algorithms may foster irrational and unjustified disparities in treatment. The third area of analysis is intellectual property, i.e. the establishment of a system of rules aimed at defining who owns the “creations” of artificial intelligence. This learning path follows the regulatory development process in the field of algorithm law, taking as its primary reference points two European Union legislative acts: the Digital Services Act (EU Regulation 2022/2065) and, above all, the Artificial Intelligence Act (EU Regulation 2024/1689), during the process of its content being incorporated into the Italian legal system through Law No. 132 of 23 September 2025, containing “Provisions and delegations to the Government on artificial intelligence”.