Governing copyright risks in generative music systems: The Suno case in the EU, US, and China

Consumer generative music platforms such as Suno strain copyright governance because training data are opaque, outputs can imitate protected recordings or performers, and attribution remains difficult at scale. This article treats Suno-type services as socio-technical pipelines from training to generation to distribution and develops a systemic risk governance framework to map where risks arise, how prevention duties are allocated, and where evidentiary burdens fall. Using a comparative qualitative case study, it analyses disputes, regulatory instruments, and platform controls in the United States, the European Union, and China. The findings identify three dominant governance configurations addressing the same structural problem: a comparatively adjudicatory configuration in the United States, a comparatively precautionary configuration in the European Union, and an agile-statist configuration in China. Despite these differences, all three are constrained by weak attribution infrastructure. The article advances the concept of informational conditions as a distinct object of legal analysis and shows a partial convergence toward platform-level governance at the distribution stage. It further suggests that collectively administered licensing frameworks may offer a more transparent and durable alternative to reliance on confidential bilateral settlement alone.