生成式人工智能输出虚假信息的侵权责任研究——以全生命周期理论为视角

Research on Tort Liability for Generative AI Producing False Information:from the Perspective of the Products Life Cycle Theory

弱人工智能时代下,生成式人工智能输出虚假信息侵权问题具有责任主体弥散化、侵权行为复杂化以及损害后果多样化等突出特征,而当前我国的单一提供者责任体系存在不足,有必要引入全生命周期理论,重视系统化思维,为传统侵权法提供一个全面审视人工智能技术的新框架。以此为基础,在责任主体方面,应以研发者与部署者为主导,使用者与传播者共同参与分配;在责任类型方面,应当将其纳入智力产品责任范畴;在归责原则方面,以过错原则为主,在涉及生命健康等重大利益的特殊领域适用无过错原则,符合产业发展的长远目标;在免责事由方面,避风港原则并无适用余地,应当构建起一般免责与具体免责相结合的免责事由体系。

In the era of weak artificial intelligence,the phenomenon of generative AI producing false information is characterized by dispersion of liability subjects,complexity of infringements and diversified damage consequences.The current single-provider liability system in China is insufficient.It is necessary to introduce the products life cycle theory,which emphasizes systematic thinking,and provides a new framework for traditional tort law to comprehensively examine AI technology.Based on this,in terms of liability subjects,developers and deployers should take the lead,with users and disseminators participating together;in terms of liability types,it should be included in the category of intellectual product liability;in terms of the principle of liability,the principle of fault liability should be the main focus,applying the strict liability principle in special areas involving significant interests such as life and health,aligning with the long-term goals of industrial development;in terms of exemption reasons,the safe harbor rule can't be applied,and a system combining general and specific exemptions should be constructed.