Robot Creativity: An Incentive-Based Neighboring Rights Approach

Abstract

Today texts, paintings and songs need no longer be the result of human creativity. Advanced artificial intelligence (AI) systems are capable of generating creations that can hardly be distinguished from those of authors of flesh and blood. This development raises the question whether AI-generated works could be eligible for copyright protection. In the following analysis, we explore this question. After a discussion of the traditional copyright requirement of human creativity, the rationales underlying copyright protection – in particular the utilitarian incentive theory – will serve as a compass to decide on the grant of protection and delineate the scope of exclusive rights. In addition, the analysis will address the question who the owner of protected AI creations should be. Finally, the discussion of pros and cons of protection will be placed in the broader context of competing policy goals and legal obligations, such as the prospect of enriching the public domain and the question of liability for AI creations that infringe the rights of third parties.

Keywords: EU copyright law, incentive theory, originality test, creative choices, artificial intelligence, algorithmic creation, machine learning, deep learning, public domain enrichment, liability for infringement, AI-assisted works, AI-generated works, disruptive effect on human creativity