EchoChamber; Jan20 - April21
[2021]
The above images are stills from a video of the artwork.
The project is the act of a closed loop in the form of an endlessly self-referential bordering-nihilistic endo-system; a sort of corporeal echo chamber, or technological body whose endocrine system is composed of trapped text, images, and voice. The work is composed of a video with accompanying audio and text produced from live code; all of which have been produced with different forms of AI and machine learning models wherein all the inputted data comes from the period of January 2020 to April 2021, a period early in the covid-19 pandemic of lockdowns and upsurge of digital media culture.
The images are compiled from over 2,000 photos, screenshots, and media from the artist's phone during the period. These images were then used to train a GAN machine learning model and discriminated against images of landscapes. Over 50,000 words of text were recorded from written journals during the period and used to train a predictive text model which produces live text - in a similar style and subject matter to the source material - during the work's installation. The audio utilised recordings of the artist speaking to create a vocal clone which produced distorted audio speaking lines from the code.