The emergence of Artificial General Intelligence(AGI)systems with autonomous decision-making capacity has plunged the efficacy of traditional criminal responsibility frameworks predicated upon the attribution of acts to a human agent into crisis.The aim of this research is to elucidate thenecessity of a paradigm shift from the conventional human-centric model of criminal responsibility and to propose a novel framework for recognizing"Electronic Criminal Legal Personality"thereby addressing the vacuum created by the independent functioning of super-intelligent systems within the criminal justice system.This article advocates for granting such systems"electronic criminal legal personality"as opposed to mere civil personality,since only through this approach can the void arising from three fundamental characteristics of AGI namely"autonomy in the selection of means"algorithmic opacity"and"the absence of human mens rea" be filled Employing an analytical-descriptive and comparative approach,the present study draws upon the criminal liability of legal persons as both a precedent and a theoretical foundation for this paradigmatic transition:just as a corporation can be the subject of criminal sanctions,so too can a super-intelligent system possess the capacity to bear punishment,albeit with a distinct nature and criteriaThe paper's central argument is that the mental element of the crime,in this context,is redefined not as internal intent but as a form of"presumed algorithmic fault," established upon the system's failure to adhere to safety standards mandated by a statutory regulatory body.The findings confirm that this model,complemented by a"mandatory guarantee fund"for victimcompensation,constitutes not a mere theoretical choice but a practical imperative for preserving the efficacy of the criminal justice system in the age of super-intelligent systems.