人工智能的刑法规制  

Regulation of Criminal Law for Artificial Intelligence

作　　者：彭文华 PENG Wen-hua(School of Criminal Justice,Shanghai University of Political Science and Law,Shanghai 201701,China)

机构地区：[1]上海政法学院刑事司法学院

出　　处：《现代法学》2019年第5期135-151,共17页Modern Law Science

摘　　要：关于人工智能可否构成犯罪,国外主要存在“心理要素说”“智能代理说”“法人类比说”“法定实体说”以及“当然主体说”等观点。人工智能成为犯罪主体的哲理基础是科学实证主义和道德二元论,现实条件是具备法律人格。智能代理可以成为道德代理。人工智能的道德生成进路有“实在论进路”“关系论进路”“认识论进路”等之别,“实在论进路”相对合理。处罚人工智能既能为处罚其他主体奠定基础,也能达到剥夺其犯罪能力的目的,还能产生附加价值。我国的人工智能刑法研究尚处于初始阶段,未来研究的重点在于:人工智能对刑法人类中心主义的挑战,人工智能犯罪主体的范畴、责任范围与程度,刑法如何介入对人工智能的规制以及人工智能犯罪的刑罚适用。Theories about whether artificial intelligence can constitute a crime,are mainly as follows:theory of psychological elements,intelligent agent theory,theory of comparison by legal human,legal entity theory and theory of the main body.The philosophical foundation of artificial intelligence as the subject of crime is scientific positivism and moral dualism,and its realistic condition is to have legal personality.Intelligent agents can become moral principles.The way of moral formation of AI is different from“realism approach”,“relationship theory approach”and“epistemology approach”,and“realism approach”is relatively reasonable.The punishment of AI can not only lay a foundation for punishments for other subjects,but also achieve the purpose of depriving them of criminal capacity,and can also generate additional value.The research on the criminal law of artificial intelligence in China is still in the initial stage.The focus of future research lies in the challenge of artificial intelligence to the anthropo-centrism of criminal law,the scope and extent of the subject of artificial intelligence crime;the intervention of the regulation to artificial intelligence in criminal law,and the application of the penalty of artificial intelligence crime.

关 键 词：人工智能 人工智能犯罪 智能代理 法律人格 可罚性 

分 类 号：DF61[政治法律—刑法学]