Copyright in AI-Generated Works and International Law Issues: A Quest for a Mediation-Based Solution

Abstract

The proliferation of Artificial Intelligence (AI) systems in the production of text, images, music, and software has intensified discussions at both national and international levels regarding how the concepts of "work," "authorship," and "ownership" should be interpreted within human-centric copyright regimes. This article examines issues such as whether AI-generated outputs are eligible for copyright protection, how ownership should be distributed among the user, the developer company, and/or data providers, and which legal order should apply to disputes regarding these outputs, from an international law perspective. The gaps created by the human-centric copyright approach of the Berne Convention in the AI era are evaluated; WIPO's work in the field of AI and intellectual property is compared with different regulatory trends shaping in the US, EU, and China. Furthermore, data mining exceptions, "fair use" debates, and the private law dimensions of cross-border infringement claims in the context of training AI models with copyrighted works are discussed. Finally, against the slow and complex nature of court proceedings in international copyright disputes, it is argued that mediation-based alternative dispute resolution mechanisms (especially institutional models like the WIPO Arbitration and Mediation Center) can offer an effective and flexible solution path for AI-origin disputes.

Comments