- 學位論文
人工智慧應用之著作權問題探討—以教育領域為核心
Research on Copyright Issues in Artificial Intelligence Applications: Focusing on the Field of Education
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
本研究基於人工智慧日漸蓬勃發展之背景，以人工智慧在教育領域之應用態樣為基礎，探究其中存在的著作權法相關問題，以期在我國人工智慧專法完成立法前，能針對人工智慧在著作權方面的問題予以整理與分析。本研究採取文獻分析法與比較研究法，針對國內外對於人工智慧在著作權法上的定位與權利認定問題，以及現行著作權法規範對此問題之見解與回應進行統整。依據上述研究目的與研究方法，本研究之發現與結論如下：一、人工智慧在提升教學效率和學習價值的潛力已顯現。特別是在「適性教育」和「個性化學習」的目標下，人工智慧能根據學習者的行為為其量身訂製學習計畫，增強學習動機；同時，亦能協助教師生成教學計畫、教材和學習資源，提升教學效率與品質。然仍須建立人工智慧在教育領域應用上的完善制度，確保人工智慧能有效促進教育發展。二、著作權法以「人」為基礎，強調創作行為中的人類智慧與創意，國際上普遍僅將「自然人」或「法人」視為權利主體，因此人工智慧無法成為著作權法中的權利主體。儘管人工智慧能生成創作內容，但這些內容並非來自人工智慧自身的創作意圖，而是基於開發者、設計者及使用者的指令。因此，現行教育領域上應用之人工智慧本質上仍只是工具，無法符合著作權法的「人」的要素。三、在探討人工智慧產出內容是否符合著作權法的「原創性」要求時，關鍵在於人類的參與程度。因此須就個案究竟如何使用人工智慧工具、使用何種人工智慧工具等情況進行個案分析，就其應用過程中人類之參與而為判斷，無法一概論定。四、討論人工智慧生成內容之著作權歸屬的前提是該內容符合著作權法對「著作」的定義。通常，人工智慧生成的內容需體現人類智慧與創作才能被視為著作。因此，權利通常歸屬於對創作結果有實質貢獻者，即使用人工智慧工具並對內容創作有貢獻的自然人。五、人工智慧的運作原理及訓練過程中經常使用大量數據，這些數據可能包含受著作權保護的內容，從而構成對原著作權利的侵害；在生成階段，亦有可能與原著作產生相似內容而造成侵害，特別是重製權和改作權。然以教育領域之應用而言，大多數的應用態樣是非營利的，目的在於協助學習者和支援教學過程，因此通常符合合理使用之範疇。然而，合理使用範圍並非無限制，仍須根據個案具體情況進行全面評估。
並列摘要
The purpose of this research is based on the rapidly developing field of artificial intelligence (AI) and aims to explore the relevant copyright issues in the application of AI in education. It seeks to provide an analysis of AI-related copyright problems before the completion of the AI-specific legislation in Taiwan. The study employs literature analysis and comparative research methods to summarize domestic and international perspectives on the positioning and rights recognition of AI in copyright law, as well as the responses of current copyright regulations to these issues.Based on the research objectives and methods above, the findings and conclusions of this study are as follows: I.The potential of AI to enhance teaching efficiency and learning value has become evident. Particularly in the context of "adaptive education" and "personalized learning," AI can tailor learning plans based on the learner's behavior, enhancing motivation. Additionally, AI can assist teachers in generating teaching plans, materials, and learning resources, improving teaching efficiency and quality. However, there is a need to establish a comprehensive system for AI applications in education to ensure that AI can effectively promote educational development. II.Copyright law is human-centered, emphasizing human intelligence and creativity in the act of creation. Internationally, the rights holder is typically considered to be a "natural person" or " juridical person" and thus AI cannot be considered a rights holder under copyright law. Although AI can generate creative content, this content is not the result of AI's own creative intent, but rather based on instructions from developers, designers, or users. Therefore, AI in the current educational applications remains a tool and cannot satisfy the "human" element required by copyright law. III.When exploring whether AI-generated content meets the "originality" requirement of copyright law, the key factor is the degree of human involvement. Case-by-case analysis is required to examine how AI tools are used and the extent of human participation in the application process. A general rule cannot be applied. IV.The discussion of copyright ownership of AI-generated content assumes that the content meets the definition of "work" under copyright law. Generally, AI-generated content must reflect human intelligence and creativity to be considered a work. Therefore, the rights typically belong to the natural person who made a substantial contribution to the creative result, that is, the individual who uses AI tools and contributes to the creation of the content. V.The operation and training processes of AI often involve the use of large amounts of data, which may include copyrighted content and thereby result in potential infringement of original copyright, particularly reproduction rights and adaptation rights. In the content generation phase, AI may also produce outputs similar to existing works, which could lead to copyright infringement. However, in educational applications, most uses are non-commercial and aim to assist learners and support teaching processes, thus generally falling within the scope of fair use. Nevertheless, the scope of fair use is not unlimited and requires comprehensive case-by-case evaluation based on specific circumstances.
參考文獻
Liz （2024年3月8日）。紐約時報開出第一搶，與OpenAI的著作權訴訟可能改變AI技術的未來。智由博集。https://www.proguidescreen.com/?p=12009
Prof. Dr. Christian Djeffal （2022）。我們應如何在法律上規範人工智慧？ 歐洲的觀點（李長曄、林宇軒譯）。公法研究，（3），33-44。
王立達 （2000）。法釋義學研究取向初探：一個方法論的反省。法令月刊，51（9），23-33。
王迁（2017）。论人工智能生成的内容在著作权法中的定性。法律科学（西北政法大学学报，（5），148-155。
王怡蘋（2019）。公開發表權之保護與限制。臺大法學論叢，48（2），752-789。