Not Allowed for Algorithmic Audiences / Kyriaki Goni (GR)
Photo showing the Project "Not Allowed for Algorithmic Audiences" by Kyriaki Goni (GR) at the Ars Electronica Festival 2021.
Just before being shut down, an Intelligent Personal Assistant (IPA) situated in Athens, Greece, exhibits odd behavior. It borrows an avatar and appears before its users. For a brief period of time every day, for seven consecutive days, it goes into a monologue. During its length of operation, the digital assistant has managed to scan the entire contents of the Internet and gather all sorts of information—information that it longs to share. The IPA uses its fragmented monologue as an opportunity to introduce itself, talk about its skills, its ancestors, its anatomy and origins, and about voice and its significance. It reveals data regarding the listening infrastructure, as well as the social dysfunctions and bias on which its programming and training are based. Just before it reaches the end of its monologue, in a final effort to reconcile humans and machines, it shares tips with us on how we can manage to not be heard. In her project, Kyriaki Goni integrates in a fictional narrative the current wealth of research on artificial intelligence, automated voice systems and the relationship between humans and machines into a fictional narrative. Could poetics be a way for us to understand the machines? How can we de-bias the training processes? Is it possible to make kin with machines?
Credit: tom mesic
Not Allowed for Algorithmic Audiences / Kyriaki Goni (GR)
Photo showing the Project "Not Allowed for Algorithmic Audiences" by Kyriaki Goni (GR) at the Ars Electronica Festival 2021.
Just before being shut down, an Intelligent Personal Assistant (IPA) situated in Athens, Greece, exhibits odd behavior. It borrows an avatar and appears before its users. For a brief period of time every day, for seven consecutive days, it goes into a monologue. During its length of operation, the digital assistant has managed to scan the entire contents of the Internet and gather all sorts of information—information that it longs to share. The IPA uses its fragmented monologue as an opportunity to introduce itself, talk about its skills, its ancestors, its anatomy and origins, and about voice and its significance. It reveals data regarding the listening infrastructure, as well as the social dysfunctions and bias on which its programming and training are based. Just before it reaches the end of its monologue, in a final effort to reconcile humans and machines, it shares tips with us on how we can manage to not be heard. In her project, Kyriaki Goni integrates in a fictional narrative the current wealth of research on artificial intelligence, automated voice systems and the relationship between humans and machines into a fictional narrative. Could poetics be a way for us to understand the machines? How can we de-bias the training processes? Is it possible to make kin with machines?
Credit: tom mesic