論文字資料探勘行為涉及的資料庫保護爭議──以科技保護措施為中心

Database Protection Issues Involving Text and Data Mining-Centering on Technical Protection Measures

在大數據（big data）時代，研究方法從過往傳統的假設導向轉為探索發現導向，透過統計學科和人工智慧（AI）領域的工具與資料庫管理技術相結合，利用大數據進行相關性研究。在電子資料庫中應用文字資料探勘技術，即是「在巨量資料中發現隱藏的知識」。在我國和美國，電子資料庫可以「編輯著作」（compilation）受著作權法保護。歐盟則採著作權和資料庫特別權（sui generis）保護雙軌制。電子資料庫業者為了因應電子資料庫內容易遭人複製與散布之侵權問題，除了積極透過訴諸司法來維護電子資料庫的著作權或資料庫特別權之外，亦安裝科技保護措施來強化電子資料庫的保護。科技保護措施對電子資料庫的保護衍生出一個急需解決的棘手問題，即學術研究者基於正當學術研究目的，規避科技保護措施以利用向公眾提供的電子資料庫進行文字資料探勘研究的適法性。本文採用比較法學研究方法，借鏡美國數位千禧年著作權法、歐盟著作權指令（Directive 2001／29／EC）、歐盟資料庫保護指令（Directive 96／9／EC），以及2019年數位單一市場著作權指令（Directive（EU）2019／790）關於文字資料探勘相關規定，探討反規避科技保護措施條款廣泛限制電子資料庫之文字資料探勘行為的適法性問題，並且透過解析歐盟增訂文字資料探勘例外規定之背景、過程和條文，以及彙整美國法院、學界對於文字資料探勘行為涉及規避科技保護措施爭議之見解，提供本文意見作為主管機關日後修訂著作權法或著作權法第八十條之二第三項各款內容認定要點之參考。

In the era of big data, the research method has moved toward from a traditional hypothesisoriented direction to a discovery-oriented direction. Through the combination of statistics, artificial intelligence (AI) and database management technologies, researchers use big data to conduct countless correlation researches. The application of text and data mining technology in an electronic database is to discover hidden knowledge in volume of data. In Taiwan and the United States, an electronic database may be protected as a compilation under the copyright law. The EU provide a dual protection system of copyright and sui generis database right to an electronic database. Other than taking legal actions, many electronic database providers also install technological protection measures to prevent content of their electronic databases from unauthorized copy and distribution. The additional protection of technological protection measures has incurred a pressing issue that needs to be resolved immediately. That is, the legality of academic researchers to circumvent technological protection measures in order to conduct text and data mining researches on an electronic database lawfully provided to the public. This paper compares relevant anti-circumvention provisions of the U.S. Digital Millennium Copyright Act, EU Directive 2001/29/EC, Directive 96/9/EC, Directive (EU) 2019/790, and the Taiwan Copyright Act. This paper addresses an important question regarding the legality of over-restrictive anti-circumvention provisions on text and data mining conducted by researchers by exploring the legislative reasons for the newly enacted text and data mining statutory exceptions in EU. Intending to help the competent authority and courts in Taiwan to handle data-mining issues in the future, this paper proposes some possible solutions to the above-mentioned circumvent problems according to the analysis of relevant US court cases and scholars' opinions regarding the use of technological protection measures to block lawful use, fair use or non-infringing use.