摘要
为了填补数据储备在人工智能系统训练中趋于“消耗殆尽”的空缺,主权国家选取跨境流动供给的方式补充数据存量,以满足人工智能系统训练对数据的海量需求。人工智能训练数据跨境供给的规制存在区域版权法中“合理使用”认定标准模糊和数据跨境流动监管标准不一致的缺陷。规制缺陷的成因在于价值取向的分歧、法律传统的差异、技术发展与规制滞后的固有矛盾,以及国际合作机制的缺失。当前,我国亟须在尊重国际治理共识的基础上,完善著作权合理使用制度,促进数据跨境有序流动,探索具有本土特色的制度创新路径。
To address the growing depletion of data resources for training artificial intelligence(AI)systems,sovereign states have increasingly turned to cross-border data supply to meet the massive data demands of AI model training.The regulation of cross-border supply of AI training data has two deficiencies:the standard for determining“fair use”in regional copyright laws is ambiguous,and the regulatory standards for crossborder data flow are inconsistent.The causes of regulatory deficiencies lie in the fundamental differences in value orientation and legal traditions,the contradiction between technological development and regulatory lag,and the absence of international cooperation mechanisms.Under the current circumstances,it is imperative for China to refine its copyright fair use system in alignment with international governance consensus,facilitate cross-border data flows,and thereby develop an innovative pathway with localized adaptability.
作者
郭德香
佘恺宁
GUO Dexiang;SHE Kaining(School of Law,Zhengzhou University,Zhengzhou 450001,China)
出处
《华北水利水电大学学报(社会科学版)》
2026年第2期82-89,共8页
Journal of North China University of Water Resources and Electric Power (Social Science Edition)
基金
河南省社会科学规划决策咨询项目(2024JC106)。
关键词
人工智能
训练数据
跨境供给
版权保护
数据跨境流动
artificial intelligence
training data
cross-border supply
copyright protection
cross-border data flows
