Symbolic harm and digital abuse: towards a new legal framework for AI-generated CSAM

Abstract

The rapid development of artificial intelligence (AI) has presented significant ethical and legal challenges, particularly in the field of child’s rights protection. This thesis asks: To what extent can AI-generated child sexual abuse material be criminalised under current national and international legal frameworks, particularly in the absence of real children as identifiable victims? While some AI-generated content may include a depiction of a real child’s face, others are entirely synthetic creations that do not involve identifiable minors. This raises urgent questions regarding the definition of crime, harm, and victimhood within criminal law. This research focuses on examining to what extent such material can be prosecuted under existing legal frameworks and what legal interest is protected. Although no physical abuse occurs during the production of such content, unauthorized use of children’s likenesses -whether real, altered or digital synthesized- raises serious legal and ethical concerns. The thesis also describes the current state of international and national legal regulation and argues that a differentiated legally sound approach to AI generated CSAM is urgently needed. Considering the potential for the normalisation of child abuse, identity harm and structural exploitation, this research argues for stronger normative responses that prioritize the best interest of the children in the digital age. Keywords: artificial intelligence, child sexual abuse materials, deepfakes, AI-generated images, legal theories, criminal law.