Unlocking the Black Box of AI Listening Machines: Assemblages for Art, Technology and Innovation

• Main Author: Sharath Chandra Ramakrishnan
• Format: Article
• Language: Catalan
• Published: Universitat Oberta de Catalunya 2020-09-01
• Series: Artnodes
• Subjects: intelligent agents; computational art and design; machine listening; socio-technical systems; auditory scene analysis
• Online Access: https://raco.cat/index.php/Artnodes/article/view/374038

The black box of innovation in the realm of connected AI technologies renders not only their technicalities opaque but also, and more importantly, the social effects and relations that constitute their creation and mediation. This presents an opportunity for creative interventions by artists and researchers, to unveil the networked relations that are part of AI technologies, and speculate on their ontological effects. This article presents such an unpacking around an AI listening machine present today in ubiquitous devices like voice assistants and smart speakers, and incorporates computational models of machine audition. By tracing the scientific research, technical expertise, and social relations that led to our cultural adoption of AI listening machines, the article presents a socio-technical assemblage within which these machines operate. At the same time, the article reveals various contexts for artists as well as innovation researchers to engage with the socio-technical complexity of AI listening machines, by sharing some instances of creative and artistic interventions that have attempted to unveil the nature of their assemblages.