Abstract

Authors: Desara Dushi, Nertil Berdufi, Anastasia Karagianni 

Generative AI has only gained public prominence in the past two years, yet instances of AI-generated CSAM videos have already been observed. It can be foreseen that in the next five years, these videos and images will become more realistic and widespread. In the United States, the FBI is already handling its first cases involving the generation of AI CSAM. This paper employs a comprehensive legal analysis of existing EU laws, including the AI Act, the General Data Protection Regulation (GDPR), the Digital Services Act (DSA), the proposed Child Sexual Abuse Regulation (CSAR), and the Child Sexual Abuse Directive to address the critical question of whether generative AI can be effectively policed to prevent the creation of deepfakes involving children. While EU legislation is promising, it remains limited, in particular regarding the regulation of training data used by generative AI technologies. To comprehensively address AI-generated CSAM, a proactive, effective regulation and holistic approach are required, ensuring that child protection against online CSAM is integrated into the guidelines, codes of conduct, and technical standards that bring these legal instruments to life.

Read the article here.