This paper explores the intersection of generative artificial intelligence and criminal liability, focusing on the rise of deepfake-enabled fraud and the challenges of attribution in 2026. It examines how synthetic media technologies disrupt traditional evidentiary standards and complicate the identification of perpetrators in digital environments. Drawing from criminology, behavioral psychology, and international criminal law, the study highlights the risks posed by generative AI to legal accountability and the protection of victims. The analysis situates deepfake fraud within broader frameworks of cybercrime and digital victimization, emphasizing the structural limitations of current governance models. By integrating interdisciplinary perspectives, the paper proposes strategies for strengthening attribution mechanisms, enhancing evidentiary reliability, and ensuring that criminal liability adapts to the evolving technological landscape. Ultimately, it argues for a reconfiguration of legal and psychological approaches to address the sociotechnical complexities of generative AI, offering pathways toward more resilient systems of justice in the digital age.
Sergio Pommier Gallo (Fri,) studied this question.