Feasibility of Redefining the Mental Element and the Problem of Attribution in Criminal Law Related to Artificial Intelligence

Abstract

The emergence and expansion of artificial intelligence–based technologies with partial or full autonomy have posed unprecedented challenges to the traditional foundations of criminal law. The core of this challenge lies in the difficulty of applying classical concepts of the “mental element” (intent, knowledge, negligence) and the “attribution of criminal responsibility” to a non-human entity lacking consciousness and free will. This article, using a descriptive–analytical method and relying on library and documentary sources, examines the feasibility of reinterpreting and redefining these concepts in light of the nature of artificial intelligence. The findings indicate that the Iranian criminal law system, similar to many other legal systems, is inherently “person-centered” and faces serious theoretical and practical obstacles in recognizing the direct criminal liability of artificial intelligence, such as the inability of such entities to bear traditional punishments and the failure to fulfill the objectives of punishment, including rehabilitation and reform. Consequently, current legal approaches primarily focus on the liability of related human actors (designer, manufacturer, programmer, user) or on the development of civil liability or risk-based frameworks. Nevertheless, this article argues that, given the increasing complexity and autonomy of such systems, a functionalist and teleological interpretation may allow the concept of “algorithmic intent” or “programmed defective performance” to be analyzed as a novel form of the mental element. Furthermore, through mechanisms such as “conditional virtual legal personality” or a “hierarchical hybrid liability system,” a pathway for attributing responsibility may be established. Finally, the article proposes the necessity of enacting specific legislation, establishing specialized regulatory institutions, and introducing mandatory insurance schemes for high-risk technologies as essential infrastructures for this theoretical transformation.