Talks · Music and AI
What the Label Misses: A Hybrid Multidimensional Taxonomy for Analyzing Heterogeneous Uses of GenAI in Popular Music
AIMC 2026, Conference on AI Music Creativity. TU Berlin and Staatliches Institut für Musikforschung, Berlin, September 2026.
Presented as a poster and published in the conference proceedings. The paper proposes a way of describing precisely what a generative system contributed to a finished piece, instead of collapsing very different practices under a single label.
AI Slop: escucha suspicaz y politización de la inteligencia artificial en música popular
[AI Slop: Suspicious Listening and the Politicisation of Artificial Intelligence in Popular Music]
Latin American Congress of Aural Studies, Universidad Mayor, Santiago, June 2026.
On what happens to listening when an audience suspects a machine was involved, and how that suspicion turned political.
Cuando la IA cantó a Sudamérica: estereotipos generativos y malentendidos humanos
[When AI Sang to South America: Generative Stereotypes and Human Misunderstandings]
First Colloquium on the Anthropology of Music and Dance, Universidad Autónoma del Estado de México, Toluca, May 2026.
A reading of "Sudamérica" as a case of how models trained elsewhere imagine a region, and of what can be done with what they get wrong.
Música e inteligencia artificial: innovación y expansión creativa
[Music and Artificial Intelligence: Innovation and Creative Expansion]
IV Congress of Ethnomusicology and II TMEIS Forum, Faculty of Music, UNAM, Mexico City, October 2025.
On what these tools open up for musicians rather than what they threaten. The talk was originally developed with Bernarda Castillo out of our experience at the AI Song Contest.
The same talk, in earlier versions for general audiences:
Museo Interactivo Mirador, Expo Volcán, Santiago, 2025.
CityLab, Centro Cultural Gabriela Mistral, Santiago, 2025.
AIMC 2026
What the Label Misses:
A Hybrid Multidimensional Taxonomy for Analyzing Heterogeneous Uses of GenAI in Popular Music
Terms like "AI song" or "AI artist" get applied to work that has almost nothing in common: a track generated from one sentence and published untouched, a piece built over months out of material the composer generated, took apart and rebuilt, a voice cloned without consent. Called the same thing, they get described the same way, and whatever human work went into them disappears.
This paper proposes a way of telling them apart. It describes a finished piece along three independent dimensions: what kind of operation the system performed in the voice, the music and the lyrics; how much of the resulting material came from the machine; and how closely the person worked with the system throughout the process. I presented it as a poster at AIMC 2026 in Berlin, and it is published in the conference proceedings.
Teaching ·
Music and AI
Creación musical con IA y espacialización Dolby Atmos
[Musical Creation with AI and Dolby Atmos Spatialisation]
CRTIC, Chile · 2025 and 2026
I teach the generative AI module of this course, for which I wrote the syllabus and an assessment rubric that weighs, among other things, the creative process and the degree of human agency in the result.
Desbloquea tu creatividad musical con Inteligencia Artificial
[Unlock Your Musical Creativity with Artificial Intelligence]
A workshop I designed and have delivered three times, adapting it each time from a single session to a full course, combining short lectures, live composition in front of the group and hands-on exercises.
San Sebastián University, with ChileCreativo · June 2025, 4 session
CityLab, Centro Cultural Gabriela Mistral, Artificial· January 2025, 2 session (online)
Museo Interactivo Mirador (MIM), Expo Volcán · July 2025, 1 session
