From the first flute carved out of bone to the modern synthesiser, technology and music have always been inseparable. Composers wrote rules to generate music centuries before the first computer did so in the 1950s. But today we’re standing at a different kind of threshold. The tools no longer simply execute the ideas we bring to them: they learn, they respond, they suggest directions we hadn’t thought to take.

This series asks what that changes. Is artificial intelligence just a more sophisticated instrument, or something closer to a collaborator with its own agency? Where does the music inside these systems come from? Whose voices, whose work, whose consent? And what actually happens in the studio, beyond the press releases, when an artist sits down with these tools?

Across three thematic episodes and ten full interviews, we speak with the people building this field from the inside: the creator of Wekinator, the founder of Neutone, the duo behind seven years of nonstop AI-generated death metal, and artists working with everything from Thai fiddles to quantum circuits. Some are optimistic, some are furious, and several are both.

The series is hosted by Antònia Folguera, curator at Sónar+D; powered by S+T+ARTS, an initiative funded by the European Commission; and produced in Barcelona by the Sónar+D team, in collaboration with music producer and artist Phran.

Episode 01 - Instrument or Co-Player?
 

Is AI just a more sophisticated instrument, or something closer to a collaborator with its own agency? Artists, researchers and instrument builders describe tools that suggest, resist and surprise - and search for the vocabulary to name what they're playing with.

Episode 02 - Whose Voices, Whose Data
 

Every neural network has a past. This chapter asks who made the music inside these systems, who gave permission, and who gets paid, while it looks at the artists building consent-based alternatives, from individual licensing deals to choir-owned data trusts.
 

Episode 03 - The Burnout and the Breakthrough
 

COMING SOON

Not the myth of the studio, but the real one: the two-year project that ended in burnout, the dataset curated like a recipe, the lineage of algorithmic music that reaches back to the 17th century. What actually happens when artists sit down with these tools.


Interview 01 - Anna Xambó
 

COMING SOON

Live coder and senior lecturer at Queen Mary University of London on building MIRLCa, her tool for navigating vast Creative Commons sound archives, and why the lack of control is part of what makes it interesting.

Interview 02 - yaboihanoi ญาบอยฮานอย
 

COMING SOON

Winner of the 2022 AI Song Contest on apprenticing with Thai luthiers, circumventing MIDI to synthesise directly into Thai tuning systems, and why music is never the mp3 file.

Interview 03 - Rob Clouth
 

COMING SOON

Electronic musician and software designer on the AI project that took two years and ended in burnout, the strange appeal of English-sounding nonsense, and how AI is changing the way he decides whether music is good.

Interview 04 - Marta Verde
 

COMING SOON

Visual artist and Berklee Valencia educator on using Stable Diffusion as a black box, feeding its output back into itself, and why AI should lead to more happy accidents rather than more automation.

Interview 05 - Rebecca Fiebrink
 

COMING SOON

The creator of Wekinator on releasing machine learning tools for musicians in 2008, when neural networks were deeply uncool, and why small personal datasets often beat big ones.

Interview 06 - Thor Magnusson
 

COMING SOON

Director of the Intelligent Instruments Lab in Iceland on instruments that become co-players, algorithmic music in the 17th century, and why prompting a song is a simulation of music rather than music-making.

Interview 07 - Ben Cantil
 

COMING SOON

Co-founder of Datamind Audio on licensing sounds directly from artists instead of labels, curating a dataset as a creative act, and why software is a form of human expression.

Interview 08 - Cris Checa
 

COMING SOON

Technologist and Sónar+D curator on making music like she's in a lab, why she resisted AI at first, and the difference between democratizing creation and democratizing experimentation.

Interview 09 - CJ Carr (Dadabots)
 

COMING SOON

The duo behind seven years of nonstop AI-generated death metal on open source, de-skilling, quantum synthesisers, and the question nobody at NeurIPS was asking: how do we make music harder?

Interview 10 - Nao Tokui (Neutone)
 

COMING SOON

Founder of Neutone on training your own models instead of Silicon Valley's, AI as alien rather than artificial intelligence, and why he misses the uncanny sounds of older, worse technology.