| Abstract | Historically, rights holders have pursued formal litigation to remedy copyright infringement. Although these legal proceedings can often be complex and lengthy, they generally occur after a song’s release. However, Universal Music Group's (UMG) recent introduction of five AI Agent Patents signifies a shift toward addressing copyright issues in real time, before a work is released, potentially eliminating the need for later legal action. These AI agents act like a ‘spy in the cab’ within various software applications, starting with Generative AI and potentially extending to other platforms and Digital Audio Workstations. Therefore, they act as extensions of an algorithmic techno-judicial process within the civil domain with input problem issues. Beyond merely identifying violations, these agents will assess required licensing types, evaluate potential human approval for usage, and ensure compliance with the artist's moral rights. The ‘spy in the cab’ metaphor highlights the agents’ role in the production process. UMG's proactive and potentially controversial approach raises critical questions about music production. Traditionally, record labels have exercised relatively loose control over creative processes, focusing instead on controlling markets and distribution. UMG's new strategy suggests that contemporary music production inherently involves some form of rights infringement, particularly issues surrounding Generative AI. This strategy gives UMG the opportunity to intercept these activities earlier in the creative pipeline, before the music enters the marketplace. In this presentation, I will discuss two use cases that highlight the complexities of UMG’s AI agent system and the so-called input problem. One case shows how major record labels, like UMG, transform royalty-free samples into exclusive rights, acting as a form of enclosure. While UMG's AI agents aim to enforce copyright, this approach also has issues for the future of music production and may inhibit the growth of new, culturally rich musical landscapes. |
|---|