Lecture-Performance
Under the scorching sun on the top of the melting AI iceberg, Grimes peacefully reads The Communist Manifesto to David Guetta deepfaking Eminem’s vocals. Google guru Ray Kurzweil floats by on a drifting ice floe. Kurzweil is waving hello and continuing a lively conversation with the AI-reincarnated double of his deceased father. Only distant twitter of stochastic parrots haphazardly stitching together sequences of linguistic forms interferes with the sonic idyll. But what lies beneath this pastoral landscape? What secrets hide in the murky waters of machine listening? And most important—whose voice will be heard and who is overwhelmed by the deep sea of data?
In Machine Mourning, an interactive sonic-focused experience, based on sound datasets trained with artists’ own voices, Arwina Afsharnejad and Daria Kozlova critically examine the entanglement between internet culture, pop music, digital media, military technologies, surveillance apparatus and audio data extractivism. The sci-fi plot of Machine Mourning unfolds at the opera premiere staged in the immersive digital environment simulating a melting glacier. Investigating the silent violence of extractive listening, Arwina and Daria attempt to raise concerns about the shift from comprehension to operation and explore its implications for the subjectivity of human and non-human agents.
The interdisciplinary research, which formed the basis of Machine Mourning, traces the trajectory of machine listening development from an early history marked by class inequality and bias of a different nature to current extractivist practices characterized by tech behemoths’ ambitions to compute every conceivable sound. It analyses the artistic and scholarly practices of the Forensic Architecture group, Hito Steyerl, Machine Listening collective, Mark Andrejevic and Timnit Gebru, among others.
In the lecture-performance, Arwina and Daria will share details of the research, the latest updates of the Machine Mourning development, and invite the audience to imagine collectively the depth of extractive listening pushing back from the bottom together in order to emerge beyond the void of overpowering machine learning.