Criminal Law Evaluation of Generative Artificial IntelligenceParticipating in Deception

Generative artificial intelligence (GenAI) participates in deception by changing the technical path of traditional fraud, strengthening the ability to realize deception, and bringing impact on the understanding of the constitutive elements of existing crimes. This paper focuses on two types of specific behavior: one is to use the self-learning function of generative AI for deep forgery, so that AI can participate in traditional fraud as an auxiliary tool. The second is the misleading deception of recommendation, retrieval or question and answer. The research believes that generative artificial intelligence deception is only the criminal "tool" for the perpetrator to implement the crime of fraud and other traditional crimes. In terms of regulation path, the first is the criminal law level, which distinguishes the principal offender of fraud crime, the accomplice of fraud crime and the crime of helping trust and other different evaluation paths; Second, at the governance level, people should promote it in four directions: data governance, technology governance, transparent governance and platform responsibility. The significance of this paper is to provide a more operational analysis and implementation framework for the criminal law identification and multiple governance of generative AI deception.