Cohorte Pratiques Sonores | Chantier IA

As part of Chantier IA, Sporobole welcomes a cohort of three artists – Amélie Laurence Fortin, Maurice Jones and Estelle Schorpp – for a six-week creative residency dedicated to the exploration of generative AI technologies. With backgrounds in music, sound art, podcasting, radio, spoken word and performance poetry, these artists will work in parallel to develop their respective projects. Accompanied by one of our digital technology developers, they will have the opportunity to experiment, share their knowledge and design prototypes or sequences of works. Oriented towards research and experimentation, this residency is not necessarily aimed at producing finalized works, but rather at opening up new artistic and technological perspectives.

Une composition numérique représentant un diamant aux multiples facettes. Le diamant, qui occupe une grande partie de l'image, intègre des éléments de visage humain, dont des lèvres bleu vif, un nez et des yeux en cristal. En arrière-plan, des textures abstraites en noir et blanc évoquent des motifs organiques ou numériques. L'image dégage une impression futuriste et surréaliste.

Amélie Laurence Fortin

universal Grammar

With the advent of digital bioacoustics—digital recorders that allow scientists to capture animal sounds continuously without the disruptions caused by human observers—we are witnessing an unprecedented surge in data. This influx has recently enabled AI to develop natural language processing algorithms capable of detecting patterns in non-human communication, bringing us closer than ever to Noam Chomsky’s concept of universal grammar. Research in this area is fundamentally reshaping the anthropocentric methodologies currently used to assess animal intelligence, with the potential to challenge our place within the animal kingdom and redefine how we interact with other living beings.

During the Chantier IA 2 – Cohort – Sound Practices residency, the artist will explore AI-driven advances in translation and communication, with the goal of collaborating with one of the four leading university research groups studying insect and animal language. By engaging with their databases and mathematical approaches to algorithm design, they aim to gain insight into the methodological challenges of translation. Ultimately, the use of generative tools will allow them to reflect on evolving concepts of translation and decoding while building a new sonic archive—one that will serve as the foundation for works exploring communication with the non-human.

📷 Digital composition. Courtesy of the artist.

Une scène naturelle en milieu forestier. On y voit un tronc d'arbre mort entouré de branches moussues et de feuilles mortes au sol. Une enregistreuse audio portative avec un micro est placée parmi les branches, suggérant une captation sonore de l'environnement naturel. L'eau stagnante et la mousse verte sur le sol renforcent l'idée d'une immersion sonore en pleine nature.

Maurice Jones

settler-native-slave

settler-native-slave is a research-creation project that explores the intertwined paths of Black and Indigenous liberation through generative, spatialized sound. Scholars Eve Tuck and K. Wayne Yang describe settler colonialism as an “entangled triad structure of settler-native-slave.” However, historical narratives about Indigenous and Black peoples often focus solely on their interactions with European settlers. This oversimplified perspective distorts history, obscuring the complex cultures, politics, and relationships that developed between Indigenous and Black communities across the Americas. As a result, it limits the possibilities for collective, liberatory futures.

Building on B. Coleman’s Technologies of the Surround, my latest work, feral.ai, applies the histories of marronage—particularly the stories of runaway slaves who escaped into the “wild” of the Great Dismal Swamp to form liberated communities—to machine learning. This approach seeks to unmoor technology, transforming it into a generative tool for liberation. While the French etymology of the word maroon conveys a sense of abandonment without resources, the so-called wild was far from empty. Indigenous communities had inhabited the swamps for millennia, leaving behind material artifacts and place-based knowledges that not only aided Maroon survival but also fostered their thriving.

settler-native-slave is the final installment in a trilogy of generative, spatial sound works. Soundscapes of an Earthly Community envisioned alternative futures of Black liberation. feral.ai explored the role of unmoored technologies in realizing those futures. Now, settler-native-slave turns its attention to the journey itself—tracing the entangled histories of Indigenous and Black communities and asking: Who do we encounter along the way, and how might these connections shape collective, liberatory futures?

📷 Feral.ai, 2024, Maurice Jones. Courtesy of the artist.

Un dessin ancien de style manuscrit, comportant des schémas mécaniques détaillés. On y voit des croquis d'oiseaux posés sur une structure mécanique composée de poulies, de tuyaux et de dispositifs circulaires, avec des annotations manuscrites en grec ancien. Le dessin rappelle des croquis de machines automates inspirés des inventions de l'Antiquité.

Estelle Schorpp

A Stochastic Spring

A Stochastic Spring is a research-creation project exploring bird songs generated by artificial intelligence.

In 1962, Rachel Carson predicted a silent spring—a world where birds are disappearing. Yet today, our soundscapes remain filled with bird songs, mediated by sonic reproduction technologies. From Ludwig Koch’s first phonograph recording of a bird in 1889 to Spotify playlists, bird song has achieved a kind of ubiquity.

While field recording, both as a cultural practice and a scientific method, has transformed our relationship with natural sounds, artificial intelligence (AI) introduces new possibilities, new questions, and a new way of engaging with the sonic environment.

In the era of the Capitalocene and the ongoing ecological crisis, how do AI tools reshape our perception of bird songs—and, more broadly, our relationship with the sounds of nature?

AI is often discussed in relation to human intelligence but rarely compared to animal intelligence. Yet birds share intriguing similarities with machines—so much so that Jacob Smith coined the term bird media in his book Eco-Sonic Media. Bird songs form a complex communication system akin to language, with some species capable of learning their vocalizations and mimicking their acoustic surroundings. Others, like canaries, have been trained to perform—once serving as sources of domestic musical entertainment before the invention of the phonograph.With this in mind, A Stochastic Spring explores several conceptual directions: AI imitating bird songs, AI conversing with birds, AI inventing new bird species, and AI recreating the songs of extinct birds.

📷 Heron of Alexandria, Pneumatica 1,16: technical drawing of a hydraulic device with artificial birdsong, Venice Manuscript, Biblioteca Marciana, Gr. 516, ff. 172v. This 13th-century codex is the earliest known text of Heron of Alexandria’s Pneumatica.

* Please note that the following text is an unedited automatic translation and is intended for accessibility purposes only.