Schrödinger's Crime: AI-generated Child Sexual Abuse Material as a Victimless Offense

Abstract

The emergence of fully AI-generated Child Sexual Abuse Material (CSAM) presents a novel and complex challenge to European Union (EU) legal frameworks. This thesis critically examines the adequacy of existing and proposed EU legislation in addressing AI-generated CSAM, with a focus on the 2024 Proposal for a Revised Directive on combating child sexual abuse. By analyzing the legal definitions of CSAM, the principle of harm, and the discretion afforded to EU Member States in criminalizing virtual CSAM, the study evaluates the implications of treating fully AI-generated CSAM as a criminal offense. Key challenges include the absence of direct victims, difficulties in distinguishing AI-generated content from real imagery, and the risks of overcriminalization. Through a doctrinal analysis of legal principles and legislative texts, this thesis seeks to clarify whether the criminalization of fully AI-generated CSAM can be justified within the boundaries of established EU legal norms, while also addressing practical enforcement concerns. The research concludes with recommendations on harmonizing EU laws to effectively regulate this emerging issue.