生成式人工智能侵权的法律规制研究

Research on Legal Regulation of Generative Artificial Intelligence Infringement

摘要: 生成式人工智能的深度应用在推动数字经济发展的同时，也对传统侵权规制体系带来了新的挑战。本文系统梳理了生成式人工智能在数据训练阶段的著作权侵权、内容生成过程中的人格权侵害以及深度伪造技术衍生的多元危害，并对当前法律规制在责任主体认定、因果关系证明及归责原则适用等方面面临的多重困境进行分析。在结合我国产业发展需求的基础上，提出构建多层级归责体系、完善因果关系认证的技术化路径、引入公证证据保全机制及推动多元主体协同治理等建议，以期为技术创新与权益保护的平衡提供理论支撑与参考。

Abstract: Generative AI, while advancing the digital economy, poses new challenges to traditional tort liability systems. This paper reviews copyright infringements in data training, personality rights violations in content generation, and harms from deepfakes, and analyzes dilemmas in identifying liable parties, proving causation, and applying imputation rules. Considering China’s industrial needs, it proposes multi-tiered imputation, technical causation verification, notarized evidence preservation, and multi‑stakeholder governance, aiming to balance innovation and rights protection.