SSC Workshop with Alexandra Supper and Yannis Patoukas

Choreographies of AI Voices: The discursive construction of “ethical” applications of AI in country music and broadcast journalism

In this talk, Alexandra Supper will analyse media discourses surrounding two different cases of AI-assisted cultural production. Both cases are framed as “ethical” applications of AI, and both cases feature AI-generated models of human voices as their centerpiece: that of American country music singer Randy Travis and of British broadcast journalist Michael Parkinson.

In spring 2024, Randy Travis has released his first new song in over a decade, after having previously been left unable to speak or sing in the aftermath of a stroke. Shortly after the release of the song, it was revealed that the song was made possible by an intricate interplay between an AI-generated model of Randy Travis’ voice and a close- knit community of human professionals with a long-standing working relationship with Travis, who himself was seen to not only consent, but also actively contribute to the making of the new song. I will compare the media discourse surrounding the release of this new song with that of another recent case that prominently features an AI-generated voice: the release of the new podcast series “Virtually Parkinson” (announced in late 2024, with episodes being launched from January 2025 onwards). In this podcast, an AI which promises to reconstruct the voice as well as interviewing style of the late Michael Parkinson, interviews various celebrities. Also as part of the podcast, the interviewees are invited to debrief about their experience of being interviewed by a generative AI in conversation with the producers of the podcast, including the son of the deceased interviewer.

The two cases show notable differences, but also some similarities, as both are presented by their makers as best-practice examples of the “ethical use” of AI for cultural production. After all, both models have been trained on data that have been provided by the legal rights-holders; both take seriously questions of consent; both are embedded in intricate, human-led working practices and promise not to threaten the livelihood of human professionals. Through a close analysis of the media coverage of the roll-out of the song and the podcast, Supper demonstrates how such a construction – to varying degrees of success and controversy – of ethical uses of AI is achieved through a careful “choreography” (employing the concept developed by Charis Thompson, 2005). Drawing on scholarship from the fields of STS, sound and music studies, disability studies and critical data studies, she addresses the cultural implications of the process by which a controversial new technology is being rendered as harmless and familiar through these discursive practices.

Studios Improvised: From 1960s experimental rock to electroacoustic music practice

This research project explores the notion of studio improvisation both as a theoretical framework and a compositional strategy, drawing from two seemingly distinct but historically and aesthetically intersecting musical practices: 1960s experimental rock and electroacoustic music. While both traditions approached the recording studio not merely as a site of documentation but as an instrument in itself, their studio practices share significant overlaps in terms of production techniques, sound manipulation, and creative workflows.

Rather than aiming for historical reconstruction, the study critically engages with selected techniques - such as tape manipulation, voltage control techniques, real time mixing, unconventional mic placement, multi-track collage, etc - as creative constraints for new work. The aim is to articulate a practice-led understanding of studio improvisation that highlights its artistic value in an era where preset-based and algorithmically generated production dominates. In doing so, the project reframes the studio not as a space of product-driven music creation but as a dynamic site of ad-hoc, real-time compositional invention.

Bio
Yannis Patoukas is an electroacoustic music composer and improviser currently based in Rotterdam (NL). He holds master’s degrees in musicology and music education (Aristotle University of Thessaloniki, 2014) and in electroacoustic music composition (Institute of Sonology, Royal Conservatoire of The Hague, 2019). He also attended the one-year Sonology course in 2016. He is currently a PhD candidate at Leiden University (ACPA) through the “docARTES” doctoral programme (Orpheus Institute, Gent). Since 2020, he is a research associate at the Institute of Sonology, where he has taught courses on analogue studio techniques and composition. As an electric guitar improviser, he performs regularly with the Sonology Electroacoustic Ensemble and collaborates with various improvising musicians both in the Netherlands and internationally. When inside a recording studio, he usually follows a more improvisational workflow; welcoming mistakes, trusting random or open processes and always making the most of the physicality of any device in the vicinity. He is attracted by the “technical limitations” found in analogue technology and enjoys the performability of tape machines, mixing desks and voltage-controlled modular systems alongside other devices or instruments.