Artificial intelligence, a “dangerous thing”?

Abstract

The present paper seeks to examine the feasibility of characterizing Artificial Intelligence (AI) as a “dangerous thing” within the conceptual framework of civil liability in Colombia. It contends that the traditional understandings of “thing” and “dangerousness” developed in Colombian doctrine and jurisprudence are inadequate to address the complex and unprecedented features of this technological phenomenon. The analysis further considers how the European Union’s classification of AI as “high-risk” introduces regulatory elements that depart from the classical foundations of civil liability. Accordingly, the paper concludes that these conceptual categories must evolve in tandem with the practical and theoretical developments surrounding AI, with the aim of achieving a coherent and necessary regulatory convergence in this field.