人工智慧生成內容與著作權：以人工智慧生成內容是否能取得著作權利保護與人工智慧系統訓練過程是否成立合理使用為核心

Copyright Issues of Artificial Intelligence Generated Content

人工智慧系統有不同的類型，其中生成式人工智慧（Generative AI）乃是從輸入的資料中進行機器學習，始能依據已知之資料內容，依賴演算法重建或創建文本、圖像、影音、甚至3D立體模型等內容，從AI系統學習到生成內容階段，均與著作權法高度相關。首先，輸入AI系統的資料可能是受到著作權權保護的客體。因而引發AI訓練過程是否侵害著作權的疑慮。其次，生成式人工智慧所生成之內容，是否可比照人類創作成果而受著作權保護，亦是深值探究的著作權法問題。
有鑑於此，本文擬針對生成式人工智慧系統衍生的著作權議題，進行探討。並將問題聚焦於人工智慧生成內容是否得為著作權利保護客體、以及生成式人工智慧訓練過程是否涉及著作權侵權。研究方法則採取文獻研究及比較法研究，由美國、歐盟的法規政策，分析生成式人工智慧系統著作權爭議的原理原則，同時對照反思我國現行法令、實務與學說見解，試圖提出未來處理生成式人工智慧相關著作權問題的參考方向。

Artificial intelligence (AI) systems encompass various types, among which generative AI is a prominent category. Generative AI employs machine learning techniques to analyze input data, enabling it to reconstruct or create texts, images, audio-visual content, and even 3D models based on known data through algorithmic processes. From the training of AI systems to the generation of content, Generative AI is intricately connected to copyright law. First, the data inputted into AI systems may constitute copyright-protected subject matter, raising concerns about whether the AI training process infringes copyright. Second, whether the content generated by generative AI can be afforded copyright protection akin to human-created works is a fundamental issue warranting in-depth exploration under copyright law.
In light of these considerations, this paper aims to examine the copyright issues arising from generative AI systems. The discussion will focus on two key questions: whether AI-generated content qualifies as a copyrightable subject matter and whether the training process of generative AI involves copyright infringement. The research adopts a methodology combining literature review and comparative legal analysis. By analyzing regulatory frameworks and policy approaches in the United States and the European Union, this study seeks to elucidate the principles underlying copyright disputes related to generative AI systems. Concurrently, it reflects on the current legal provisions, practices, and scholarly perspectives in the domestic context, with the goal of proposing reference points for addressing copyright issues associated with generative AI in the future.