生成型人工智能的刑事风险与防治策略——以ChatGPT为例

Criminal Risk and Prevention Strategy of Generative Artificial Intelligence--Take ChatGPT as an Example

生成型人工智能在通过程序生成具有创造性的新内容方面实现了新的突破,同时蕴含诸多的刑事风险。生成型人工智能的刑事风险包括不当获取数据信息的刑事风险、生成违法犯罪信息的刑事风险、使用者不当利用生成内容的刑事风险。规制生成型人工智能的刑事风险时,刑法应有所为有所不为,始终保持内在的谦抑,坚守不阻碍科技发展的底线。为此,应采用分级治理策略:通过生成型人工智能研发企业刑事合规计划,探索生成型人工智能所涉违法犯罪的预防性治理模式;加强科技监管手段在行政监管手段中的运用比重,通过有效的行政规制手段降低生成型人工智能刑事风险现实化的可能性;当刑事合规计划与行政监管手段失灵时,才可考虑采用轻重程度适当的刑法规制手段。实施分级治理策略,有利于督促生成型人工智能研发企业从源头上降低风险,并充分发挥行政监管与行政前置法“过滤”犯罪的作用,实现遏制技术风险和促进科技发展的双重目的。

Generative artificial intelligence has achieved a new breakthrough in the process of generating creative new content,while containing many criminal risks.The criminal risk of generative artificial intelligence refers to the criminal risk involved in the improper acquisition of data information in large-scale language learning and intensive training,the generation of illegal and criminal information when creating content through programs,and the improper use of content generated by generative artificial intelligence by users.Criminal law should always adhere to the bottom line of not hindering the development of science and technology,adhere to the principle of modesty of criminal law,and practice the strategy of hierarchical governance when dealing with the criminal risks of generative artificial intelligence.That is to say,it is necessary to develop enterprise criminal compliance plan through generative artificial intelligence and explore the preventive governance mode of crimes involved in generative artificial intelligence.Strengthen the proportion of scientific and technological supervision means in administrative supervision means,improve the intensity of administrative supervision,and effectively reduce the possibility of realizing the criminal risk of generative artificial intelligence through administrative regulation means.When the administrative pre-legal regulation is invalid and the act meets the standard of criminal illegality,the appropriate means of criminal law regulation can be considered.The application of hierarchical governance strategy is conducive to urging generative artificial intelligence R&D enterprises to reduce risks from the source,to give full play to the role of administrative supervision and administrative pre-law“filtering”crime,so as to achieve the dual purpose of curbing technological risks and promoting scientific and technological development.