Criminal Liability in the Age of Artificial Intelligence: From Human Agency to Machina Sapiens Autonomy

Abstract

In the context of the accelerated development of artificial intelligence (AI)-based technologies, this article examines the challenges posed by the attribution of criminal responsibility both to AI systems and to the human agents who design, program, or use them. To this end, it reviews the main models of criminal imputation related to AI, distinguishing between the liability that may fall on developers, manufacturers, and programmers, and that attributable to individuals who employ these technologies as instruments for the commission of criminal offenses. The article then explores the hypothesis that certain advanced forms of AI, particularly those endowed with a high degree of functional autonomy, could, in the future, be considered criminally responsible. Finally, it reflects on the nature and limits of the sanctions applicable in scenarios involving non-human artificial agency.