Overcoming Barriers to Text and Data Mining in the Era of ChatGPT: the Proposed Data Act as a Game-Changer

Abstract

<p>Article 35 of the proposed Data Act intends to free databases comprised of machine-generated data from the constraints of sui generis database protection. The provision may provide some remedy for researchers facing challenges under the text and data mining (TDM) exceptions of the Directive on Copyright in the Digital Single Market. However, the wording of Art. 35 may also raise questions and create legal uncertainties for scholars, as scientific research is not the main aim of the proposed legislation. This article argues that, even after the proposed Art. 35 of the Data Act, some databases comprised of machine-generated data may still fall within the scope of the sui generis database protection under certain circumstances. The discussion revolves, inter alia, around the concept of recorded data in the context of the ‘obtaining-creating dichotomy’, the use of mixed databases and the notion of derived or inferred data, and the issue of researchers’ access to machine-generated data. The findings of this article are intended to offer guidance to researchers using Artificial Intelligence tools, such as ChatGPT, to mine databases on how they may effectively avoid a potential infringement of the sui generis database right. The paper hopes to encourage new changes in the EU regulation on TDM that could create a more balanced and research-friendly framework.<br></p>