AI-Driven Technology and Its Challenge to the Law

Maurizio Ferraris (University of Turin), Gianmaria Ajani

AI-driven technologies are challenging lawmakers’ aptitude to regulate social needs. Shortfalls in the legal regime of new technologies can be traced by looking at different factors: inertia or delay in adopting regulations, an increased diversity between the EU favor for regulation and the US unregulated approach, lack of coordination between supranational policymakers and national ones, a blurred definition of boundaries between ethical and/or legal procedures.
Deficits in policymaking may activate contentious decisions by courts or regulatory authorities, such as, for instance, the recent ban on CHAT-GPT unexpectedly ruled by the Italian Privacy Authority. Who is accountable for the said inertia? Can policymakers justify themselves, calling as an excuse the unpredictability of AI advancements? Answering these questions offers only a partial response to the impasse we are facing. The matter is not only how to find a challenging balance between the protection of personal identities and recognition of overall benefits brought in by AI applications. It is, also, how to make current and future regulation fit with an established set of legal classifications. Major legal categories, established to attract and contain as a magnet the assorted aspects of economic relationships, seem not able anymore to govern the disruption introduced by AI in the production of value. Diversity between common law-based legal orders and civil law ones is already affecting global competition in the management of big data. Property rights on information boosted by big data, but also liability rules concerning risks and damages introduced by independent agents, or identification of copyright coverage for artworks made by Generative Adversarial Networks: these and other cases call for a redefinition of legal principles and categories dating back to the 19th century. Such theoretical work is needed to clarify key issues in the management of AI developments, first of all: who owns what?