This article analyzes trends and tensions around generative AI systems focusing on the contemporary music industry and unpacks the complex, at times contradictory, discourse that surrounds them. Drawing on public trade press, corporate statements, and field ethnography at international music industry trade events, we address industry questions such as profitability, regulation, and power as well as political economy issues such as ideas of fairness, crediting, and remuneration in the context of generative AI and music. Particularly, we focus on how these contradictions foster the emergence of notions of ‘responsible’ AI. In doing so, we reveal the implicit assumptions on universality of ethics currently present in this industrial field, and their cultural implications for AI development. We use critical approaches to universality in AI ethics to point to some of the pitfalls in these discourses of ‘safe and responsible’ AI in the music industry. Concurrently, we highlight how these industrial contradictions and ethical pre-assumptions are disconnected from culturally-informed perspectives on music.

When performing live, musicians have a push-and-pull dynamic with failures. They strive to maintain control by avoiding failures – by hitting the right notes and staying on the beat. Yet, to enact the spectacles of suspense in the face of live audiences, musicians also tread on the verge of messing up — they would try at something dicey knowing that such attempts could very well turn disastrous. From Auto-Tune, drum machines, to automatic beat-syncing, computational media brought forth new possibilities to configure sonic events in real-time. Despite their presumed teleological orientation toward stability, such techniques also introduced new ontological uncertainty to the live stage. These techniques supplant existing modes of corporeal control — their effects are indeterminate as they unfold in real-time, and their glitches may be irrevocable in live instances. Against this backdrop, this presentation excavates the shifting expectations of failures with the advent of computational media on the live musical stage. By zooming in on the practices of (1) live looping and (2) live coding, I show how the ontological uncertainty of real-time algorithmic processes may be repurposed as part of the live aesthetics — wherein musicians and audiences alike come to embrace the speculative possibilities of failures inherent to their ontic operations. Through the technological mediation of sonic liveness, I sketch out a reorientation toward a pragmatist ethics of computational media that rejects the cybernetics fantasies of control — and instead acknowledges failures as always already probable within such media.

The present study investigates how TikTok’s affordances—such as trends, challenges, and algorithmic visibility—mediate traditional metal subcultural dynamics while allowing for localized reinterpretations. Notably, Latin American metal fans engage with humor, national identity, and genre boundaries in ways that challenge and reinforce existing power structures within the global metal scene. The presented work is part of an ongoing online ethnographic study being conducted by the investigator, in which she analyses the content produced by 20 TikTokers from Latin America who associate with the metal and rock culture. Paying attention to issues of global and local dynamics, fandom practices, gender, genre and taste disputes, the present work makes an initial immersion in the material collected by the researcher, and delves into questions of identity construction and online performances. By examining Metaltok as a site of negotiation where digital performances interact with glocal subcultural practices, this research contributes to broader discussions on music, digital media, and gender in Latin America. It highlights how TikTok facilitates both the reproduction of metal’s historical hierarchies and the emergence of alternative narratives that reconfigure fandom and authenticity in the digital age.

Drawing on the intersections between Platform and Infrastructure Studies, this article seeks to discuss and investigate how online platforms operate as infrastructures based on which specific modes of knowledge are institutionally and collectively constituted. More specifically, our proposal is to investigate online platforms as epistemic infrastructures. Initially used to examine knowledge production in scientific laboratories, the concept of epistemic infrastructures now extends to online platforms and its data-oriented ways of tracking, measuring, and governing practices based on its interfaces, features and other materialities.

Focusing on the reconfiguration of music charts, the article explores how platformization changes the ways musical hits are quantified and made visible, reshaping power dynamics in this cultural industry. This study also highlights the role of data fandoms, transnational fan communities that use coordinated streaming and social media strategies to influence chart rankings. These fandom-driven practices reveal negotiations and tensions between industry structures and participatory cultures, and trigger changes in how platforms govern collective and coordinated epistemic practices.

Through this lens, the article frames platforms as pivotal infrastructures that not only expanded the number of metrics attached to the rankings, but have also fostered a fan culture in which the relationship with music is centrally mediated by measurable interactions. In this sense, platforms act as epistemic infrastructures in which governance of numbers is under dispute by a range of actors who can manipulate indicators, interfaces and policies and thus assert their emerging forms of knowledge.