Abstract: The rise of artificial intelligence (AI) challenges norms about technology, the Internet, and data governance. These AI tools rely on datasets or foundation models, often collected by data scraping, a process of sweeping the vast expanse of the Internet for content. A common dispute is that certain AI models use copyrighted materials to train their systems without giving attribution for the copyright holder. The question therefore centers on whether training data itself qualifies for copyright protection and if an AI model’s output crosses into copyright infringement of that data. At its core, the advancement of AI has put the notion of copyright to the test. The Bloomington school’s scholarship on the cultural commons offers theoretical insight into the history of copyright infringement, revolving around the challenge of defining boundaries. The Bloomington School’s understanding of “commons” in general, including the cultural commons specifically, is rooted in two key principles. One is that a commons is a space where people share resources. Second, many different types of institutions can be at work in a commons. These institutions include boundary rules that allow people to share resources and overcome social dilemmas. Foundational scholarship on the cultural commons, which began developing in earnest twenty-five years ago, connects the cultural commons to the public domain and intellectual property. This research was generated by critical conversations about boundaries occurring around the year 2000, particularly how the new digital revolution challenged prevailing notions about intellectual property. In effect, digital technology unleashed the human capacity to share resources, including those considered to be under the protection of the intellectual property regime. This paper argues that the concerns and innovations about sharing and boundaries in foundational texts of cultural commons research remain salient today. Thus, we trace the salient features in this early body of scholarship with our present-day case study, AI, in mind. We’ll first examine recent cases involving AI and copyright protection issues, particularly instances of infringement over the usage of AI. From there, we’ll use the theoretical framework of cultural commons to shed light on the copyright movement and the concept of copyright itself. Primary focus will be placed on how boundaries are defined within the development of copyright protection. Finally, we’ll bring these discussions together to outline a theory of copyright infringement for the AI era.
