深度伪造人工智能的刑事风险应对

Addressing the Criminal Risks of Deepfake Artificial Intelligence

深度伪造技术是被称作“生成式对抗网络”(GAN)的人工智能深度学习人类样本,将个人的外貌、声音等生物识别信息伪造合成虚假内容的人工智能技术。目前我国刑法对于深度伪造技术的规制主要存在两种模式。一是对非法获取、出售、提供个人生物信息的上游犯罪进行风险预防的前置性保护模式,二是对利用深度伪造的虚假信息实施诈骗、诽谤等下游犯罪进行事后评价的后端处置模式。但上述两种规制模式都未对非法利用深度伪造技术的行为进行刑事规范,难以规制“合法获取 + 不法利用”生物识别信息的行为,无法实现刑法的规范保护目的。因此,应当对非法利用深度伪造技术行为进行定性分析,加强对个人生物识别信息的刑法保护,将非法利用行为纳入侵犯公民个人信息罪的行为类型,弥补非法利用深度伪造技术的刑法规制空白。

Deepfake technology is artificial intelligence technology called “generative adversarial networks” (GANs) that learn from human samples and forge biometric information such as an individual’s appearance and voice into fake content. At present, there are two main modes of regulation of deepfake technology in China’s criminal law. The first is a pre-protection model for risk prevention for upstream crimes that illegally obtain, sell, and provide personal biological information, and the second is a back-end disposal model for post-event evaluation of downstream crimes such as fraud and defamation using deepfake false information. However, neither of the above two regulatory models criminally regulates the illegal use of deepfake technology, and it is difficult to regulate the behavior of “lawful acquisition + illegal use” of biometric information, and it is impossible to achieve the purpose of normative protection of the criminal law. Therefore, it is necessary to conduct a qualitative analysis of the illegal use of deepfake technology, strengthen the criminal law protection of personal biometric information, include illegal use in the type of crime of infringing on citizens’ personal information, and fill the gap in the criminal law system for illegal use of deepfake technology.