Emmie Head, “Towards an Aesthetics of Digital Emotion: Examining Emotion in AI Music Composition Process”
“Beatoven.ai uses advanced AI music generation techniques to compose unique mood-based music to suit every part of your video or podcast,” says the slogan for royalty-free generative AI Beatoven. Beatoven, like other music-making AIs, interprets a prompt from the user based on provided key terms that include duration, style, vibe, era, and occasion. After creating a track, the user simply selects the “Create Emotion” button and changes the sonic mood of the track they have generated from a table of sixteen possible options. As generative AI technologies such as this advance, it is necessary to consider who determines what these moods (or emotions) sound like and what sonic markers characterize these determinations.
Using Beatoven as a case study, I evaluate the stylistic markers used to signify mood or emotions. Employing Sara Ahmed’s analysis of affective economies, “where feelings do not reside in subjects or objects, but are produced as effects of circulation” (Ahmed, 2004), I argue that, in its capacity to use a preexisting work as a foundation and alter components to fit a user’s expectation of a mood/emotion/style, Beatoven allows its audience to circumvent circulation and access emotion within an object – that object being the prerecorded music that provides data within which information is stored to create various emotional outputs. Since affective economies exist and since emotions are shaped by contact, an AI’s contact with the prerecorded music it is trained on and the data that alters its aesthetic and affective qualities produces its emotional capacities. As generative AIs become a quick and cheaper alternative for creators who wish to use royalty-free music, it is necessary to evaluate the types of interpretations of emotion and style that these AIs produce to intimately understand who does or does not benefit from this streamlined process of musical composition.
Caleb Herrmann, “Opulent Air: Billie Eilish’s Breath and Post-War Nostalgia”
In 2021, Billie Eilish shocked fans with a new hair color: platinum blonde, a drastic change that accompanied a more comprehensive overhaul of her wardrobe. In appearances at the Met Gala and on the cover of Vogue magazine, the singer rolled out a new celebrity persona that drew upon the post-war opulence of the 1950s/60s. With references to a cluster of iconic figures like Marilyn Monroe and Grace Kelly, Eilish reactivated a period of American hegemony, unbounded economic growth, and social security for the nation’s middle-class citizens—conditions that today have largely disappeared, especially Eilish’s Gen-Z and millennial fans. Dressed in dreamy pink ball gowns, Her gesture was both melancholic and nostalgic, holding on to a social formation that has largely disappeared.
This paper uses Eilish’s opulent, post-war nostalgia as an opportunity to explore the sonic resonances between the so-called “Golden Age” of capitalism and the singer’s musical vocabulary. Although Eilish references vocalists of the 1950s/60s as influences—Peggy Lee, Frank Sinatra, Julie London—it is ultimately via breath that she most actively engages with the past. Singing in a whispered, airy tone, I consider the role of breath in recalling a time when unbounded economic growth coincided with social and political stability. As a musical resource, breath puts into play a constellation of figures associated economic growth: resource abundance and scarcity, exhaustion and recovery. In a close reading of their award-winning song “What Was I Made For?”, I show how Eilish enlists vocal and studio technique to figure breath as infinite, fabricating a musical world in which resources don’t run out. Ultimately, I argue that cultivating a vocal style that aims at un-bounding breath presents the post-war era of limitless economic growth as a solution to living in a world of stagnating growth.
Kelly Hoppenjans, “‘A Self-Replicating Pop Star?’ Grimes AI and Voicing Humanity”
In the past two years, AI voice clones have advanced rapidly, sparking controversy in pop music circles. These programs, trained on recordings of a particular singer’s voice, allow users to create new vocal tracks emulating that singer’s unique sound so convincingly that listeners struggle to differentiate between them. Enigmatic dark pop artist Grimes has enthusiastically embraced voice simulation technology, developing an AI double of her voice and inviting anyone to use it. She is optimistic that “creatively… AI can replace humans” and describes herself as a “self-replicating AI popstar.” Yet, currently, humans remain essential to this and all other AI projects, as Grimes’s “self-replication” would be impossible without the singers who transform their voices into Grimes, lending their timbres, bodies, and stylistic gestures to the assembled AI voice.
This paper explores Grimes AI from the perspective of singers who transform their voices to highlight the humans behind the AI. I asked singers who had never used it to sing through it with me, interviewed producers and artists who have created songs using Grimes AI, and experimented with my own voice through the clone. As much as singers encounter bizarre, almost totalizing transformations of their voice, they often also detect ways in which their voices are still present, resulting in a vocal hybrid that is a blend of the singer, Grimes, and the technology itself. As they negotiate these multiple identities, they experience ways in which their selves have been erased or displaced by the technology as well as moments where vestiges of their embodied sound and vocal style are still audible in the voice clone’s assemblage. As AI grows more ever-present in society, their perspectives help us understand how we can reckon with our selves, our voices, and our bodies through and despite this technology.
