文本与数据挖掘的合理使用问题研究

Research on the Fair Use of Text and Data Mining

摘要: 本文围绕文本与数据挖掘在人工智能研发过程中的著作权侵权风险展开研究，系统分析了TDM行为在复制权、改编权与信息网络传播权等方面可能构成的侵权类型。进一步探讨了将TDM纳入著作权法“合理使用”制度的理论与现实困境，包括与“作者中心主义”的冲突、利益平衡机制的破坏以及现有合理使用条款的适用局限性。在此基础上，结合欧盟、英国、美国的立法经验，提出在我国构建TDM合理使用制度的可行路径，包括引入“转换性使用”与“非表达性使用”概念、修改兜底条款为开放式一般例外条款等，以促进人工智能产业健康发展与国家创新战略的实施。

Abstract: This paper examines the copyright infringement risks associated with text and data mining in the development of artificial intelligence. It systematically analyzes potential infringements related to reproduction, adaptation, and communication to the public rights. Furthermore, it explores the theoretical and practical challenges of incorporating into the “fair use” (or exceptions and limitations) framework under copyright law, including conflicts with doctrines, disruptions to the balance of interests, and the limitations of existing fair use provisions. Drawing on legislative experiences from the EU, the UK, and the U.S., the study proposes pathways for establishing a TDM fair use system in China. Suggestions include introducing concepts such as “transformative use” and “non-expressive use”, as well as modifying the current “catch-all” clause into an open general exception clause to support the healthy growth of the AI industry and align with national innovation strategies.