人工智能侵权:产品责任制度规制的窠臼与纾解

Artificial Intelligence Infringement:The Dilemma of Product Liability System and Effective Solution

承认智能软件“产品”属性符合立法目的,且将软件集成到机器上可满足“产品”应为有形物之要求,故而运用产品责任制度规制人工智能侵权具有合理性,但也面临能否适用严格责任原则、难以证明产品“缺陷”以及难以合理确定责任承担主体的困境。对此,从比较法、经济发展和司法实践的角度审视,应当坚持严格责任原则;在产品缺陷认定标准上需要进一步建立“消费者期待标准”,同时减轻人工智能产品缺陷与损害结果因果关系的举证责任,排除人工智能产品的发展风险抗辩。为合理确定责任主体,立法应明确生产者、销售者的基本概念与主要类型,同时增设运营者作为责任主体。

Recognizing that the property of intelligent software as"product"is consistent with legislative purposes and that integrating the software into a machine satisfies the requirement that"product"should be a tangible object,it is reasonable to use the product liability system to regulate AI infringement.However,it is also difficult to apply the principle of strict liability,to prove product"defect"and to reasonably determine the subject of liability.In this regard,from the perspective of comparative law,economic development and judicial practice,we should adhere to the principle of strict liability.It is necessary to further establish"consumer expectation standard"in product defect identification standards,and at the same time reduce the burden of proof of causality between AI product defects and damage results,and eliminate the development risk defense of AI products.In order to reasonably determine the subject of liability,legislation should define the basic concepts and main types of producers and sellers,and add operators as the subject of liability.