ENGINEERED INEQUALITY: MUSICAL TAXONOMIES AND STREAMING RECOMMENDER SYSTEMS

Abstract

In the past five years, the music industry and streaming giants have embraced the success of Latin, Korean and Afrobeats music and heavily promoted their multicultural expertise and international expansion (Spotify 2021a, 2018b; Dredge 2022). However, despite academic efforts to understand streaming classification and recommendation (Seaver 2022; Maasø & Spilker 2022; Hesmondhalgh et al. 2023) it is still unclear which musical taxonomies are used by music streaming platforms. The publication of the DCMS report (2023) in the UK has brought to the fore the importance of understanding streaming software infrastructures, including how music is organised by platforms, and how recommendations and curation are automated. What kind of cultural visions, understandings, and taxonomies of music are currently hardwired into recommender systems? This paper analyses the taxonomies and the metadata standards used to code music catalogue into streaming services to argue that the current industry practices do not contribute to the international vision promoted. Using a mixed methods approach, this paper combines interface analysis of music streaming platforms, discourse analysis over PR industry materials and ethnographic fieldwork at industry conferences, and interviews with industry, public, and start-up stakeholders. In doing so, it contrasts the disparity between industry emphasis on automation, expertise and internationalisation with practices that reveal Western-centric, incoherent and error-prone approaches to catalogue. Following a postcolonial cultural economy framework (Saha 2021), I show how these underdeveloped software infrastructures contribute to the making and reproduction of culture and race, and more widely how they impact music cultures worldwide.