生成式人工智能著作权侵权损害的法理阐释及规范建构

Jurisprudential Explanation and Normative Construction of Generative Artificial Intelligence Copyright Infringement Damage

Generative artificial intelligence copyright infringement damage is difficult to regulate because of its “special jurisprudence”. The dual attributes of public and private interests of copyright breed the generative artificial intelligence in the positioning of the subject of infringement, the path of infringement liability determination, the mechanism of damages and considerable causation of the application of the dilemma of the specificity of the generative artificial intelligence copyright infringement damages and general infringement of the differences. Through the comparative study of infringement, subject fault, damage fact and causality, the specific presentation of the relevant jurisprudence is analyzed in a hierarchical manner, so as to trace back and analyze the uneven distribution of rights and interests between copyright protection and industrial development, and between individual benefits and public benefits, which is the essence of the damage derived from copyright infringement. Combined with the theory of balance of interests, for the legal regulation of generative AI infringement damages, we should clarify the scope of legal protection effectiveness and technological boundaries, introduce “special jurisprudence” in the traditional infringement damages regulation model, and construct a hierarchical infringement damages regulation model; relying on the effectiveness of the state and society’s supervision, we should set up a special system of supervision of infringement-related behaviors. It also establishes a special coordination mechanism for the regulatory system of tort-related behaviors based on the regulatory effectiveness of the state and society, creates a tort immunity mechanism for the responsible subject, and creates a synergistic governance pattern for all stakeholders. In this way, a balanced protection mechanism with both public and private attributes will be formed.

生成式人工智能著作权侵权损害难以规制的原因在于其蕴含的“特殊法理”。著作权公益和私益的双重属性孕育了生成式人工智能在侵权主体定位、侵权责任判定路径、损害赔偿承担机制及相当因果关系适用困境的特殊性，进而致使生成式人工智能著作权侵权损害与一般侵权行为存在差异。通过对侵权行为、主体过错、损害事实和因果关系等四方面的比较研究，分层次研析相关法理的具体呈现，从而溯源探析出著作权保护与产业发展、个人收益与公共受益两方面的权益分配不均是著作权侵权损害衍生的本质。结合利益平衡理论，对于生成式人工智能侵权损害的法律规制，应厘清法律保护效能与技术边界范畴，在传统侵权损害规制模式中引入“特殊法理”，建构分层级侵权损害规制模式；依托国家、社会的监管效能，建立涉侵权行为监管体系专项协作配合机制；打造责任主体的侵权豁免机制，营造各利益攸关方的协同治理格局。以此形成兼具公私属性的平衡型保护机制。