Guilt, artificial intelligence: and other questions about responsibility

Abstract

This article addresses the contemporary challenges that artificial intelligence poses to traditional legal categories of attribution, especially with regard to culpability and criminal, civil, and administrative liability. It begins from the premise that certain technological and social developments outpace the law’s capacity for timely regulation, thereby requiring a hermeneutic effort that integrates liberal values, regulatory functions, and normative adaptation. The text explores the notion of conduct as a legal-technical criterion for attribution, confronting its classical connection to human volition with the new configurations assumed by autonomous agents and AI systems. It also discusses the legal possibility of attributing responsibility to AI, drawing on comparative models — such as those proposed by Gabriel Hallevy — and examining the difficulties of assimilating AI to entities like legal persons or non-human animals.