da inserire

Abstract

This article explores the historical evolution of the concept of risk and its contemporary applications in the domain of artificial intelligence. Building on Beck’s notions of “global risk” and “risk society,”  we argue that today’s context can be described as a digital risk society, increasingly marked by uncertainty and unpredictability. The idea of systemic risk—widely discussed in complexity science and economics—emerges as a particularly suitable conceptual tool for interpreting this scenario. The centrality of risk is further evidenced by the EU’s AI Act, which adopts a risk-based approach. We examine this regulation to analyze the interaction between risk and systemic risk, highlighting the limitations of this framework.