A Small Interdisciplinary Conference About Developing Methods for Studying Algorithmic Music Ecosystems

“Fail fast, fail forward.” This phrase encapsulates Silicon Valley’s ethos and has afforded rapid technological development. It has also made these systems difficult to study: like the metaphorical “stream,” they are constantly in flux. One consequence: digital music, streaming platforms, and cloud infrastructures have been around for decades, yet scholars lack a consensus on how to study these objects. Some have taken a social anthropological approach, examining the makers of these systems (Seaver 2022, Born 2022). Others have examined the political economy of these systems’ development (Drott 2023, Scherzinger 2019). Yet others have taken the computational models themselves as a point of departure (Amoore, 2020; Kang, 2023). Most recently, developers have started applying artificial intelligence and machine learning models to their musical systems, creating a new layer of opacity for researchers. But, while a neural network’s weights and activity may remain invisible to the user, certain aspects remain accessible for study, such as the user interface, API, and some datasets.

The aim of this conference is to explore different methods for making digital music systems known. We welcome scholars across musicology, popular music studies, media studies, history of science, and critical data studies who consider the technologies that are becoming central to musical practice, such as datasets, user interfaces, and platforms. As music streaming and music generation converge in platforms like Spotify and Suno, how is music represented in the “ground truths” of these systems? What are the histories of benchmark datasets, and how have political and economic forces shaped their development?h.