The Dilemma and Solutions in Applying Criminal Law to Generative AI Fraud Crimes: The Case of Sora

Abstract

The rapid advancement of artificial intelligence systems such as Sora has brought unprecedented convenience to humanity, yet simultaneously given rise to novel forms of fraud. The emergence of such AI systems does not alter the fundamental reality that machines lack autonomous consciousness; consequently, artificial intelligence cannot constitute the principal agent in fraud offences. Fraud crimes utilising generative AI like Sora lack the requisite intent for fraud, necessitating adherence to the “disposition consciousness requirement theory”. Liability for fraud crimes should consider multiple parties—producers, users, intermediary institutions—based on their specific involvement. Addressing responsibility in AI-facilitated fraud crimes urgently requires establishing reasonable duties of care, refining relevant criminal charges and judicial interpretations for AI fraud, and constructing multi-tiered, multi-stakeholder regulatory mechanisms.