人工智能刑事责任主体否定论——基于规范与语义的考察

The Negation of Artificial Intelligence as the Subject of Criminal Responsibility:Based on Normative and Semantic Investigation

人的存在论特性决定了刑法是行为规范体系,故唯有能够理解概念和语义、能够领会刑法规范的内容和要求的主体才可能作为适格的规范接受者被视为刑事责任主体。不论是采用符号计算主义还是联结主义的系统,当前的人工智能技术都不能解决语义的理解和生成问题。根据意义的指称理论,人工智能系统因缺乏将语词与世界事物关联起来的意向能力而无法理解语义,根据意义的使用理论,人工智能系统因并未与人类分享共同的生活形式而不具有语言使用者的规范主体身份。因此,至少在可预见的将来,人工智能技术无法逾越语义鸿沟,不具有成为刑事责任主体的资格。

The ontological characteristics of human being determine that criminal law is a system of behavioral norms,so only the subjects who can understand the concepts and semantics,and the content and requirements of criminal law norms can be regarded as the subjects of criminal responsibility as qualified norm recipients.Regardless of whether the system adopts symbolic computationalism or connectionism,current artificial intelligence technology cannot resolve the problem of semantic understanding and generation.According to the referential theory of meaning,artificial intelligence systems cannot understand semantics because the lack of the intentional ability to associate words with things in the world.According to the use theory of meaning,artificial intelligence systems have no normative subject identity of the language users because they do not share a common life form with humans.Therefore,at least in the foreseeable future,artificial intelligence technology cannot cross the semantic gap and does not have the qualification to become the subject of criminal responsibility.