生成式人工智能服务提供者的侵权责任边界研究

Research on Tort Liability Boundary of Generative Artificial Intelligence Service Providers

生成式人工智能服务提供者作为生成式人工智能产品的预训练数据、优化训练数据来源筛选和控制的主体以及算法设计主体,应在生成内容侵权时承担相应责任。《生成式人工智能管理暂行办法》规定了生成式人工智能服务提供者的主体责任,但存在服务提供者的内涵不清、侵权责任认定规则不明等问题,导致生成式人工智能服务提供者权责不一,不利于生成式人工智能的深化发展。鉴于此,需要明确生成式人工智能服务提供者的具体内涵,使其与其他责任主体相区别,厘清其承担侵权责任的法律地位。在侵权责任认定路径上,可通过生成式人工智能服务提供者侵权的归责原则、过错认定、因果关系认定和免责事由等内容进行优化与扩展,达到为生成式人工智能服务提供者的侵权责任划定边界的效果。

As the subject of pre-training data of generative AI products,source screening and control of optimized training data and algorithm design,generative AI service providers should bear corresponding responsibilities when the generated content infringes.Currently,the“Measures”have stipulated the principal responsibility of generative AI service providers.Still,there are problems such as unclear connotation of service providers and unclear rules for the identification of tort liability,resulting in different rights and responsibilities of generative AI service providers,which is not conducive to the deepening development of generative AI.Given this,it is necessary to clarify the specific connotation of generative AI service providers,make them different from other liability subjects,and clarify their legal status of infringement liability.In the determination of the path of tort liability,It s reliable to optimize and expand the principle of attribution of infringement of generative AI service providers,the determination of fault,the determination of causality,and exemptions from liability to achieve the effect of delineating the boundaries of the tort liability of generative AI service providers.