Speaking in Tongues (2017)
Interactive video installation using Google Cloud Speech API, Max MSP, PyOSC, Python, ~Sigmund.
'Speaking in Tongues' is a project that explores the different types of information that are conveyed by the human voice and how machine learning is interacting with this.
Speaking in Tongues I uses various technologies to analyse in real time what the user says and classifies them in various ways (e.g. gender, adherence to linguistic standard norm etc) which triggers a visual response. Over the course of speaking the visuals shift and flicker as the voice falls into different stereotypes - more male, more female, more educated, less educated.
Speaking in Tongues II looks at what happens when machine listening gets it wrong. Echoing Joseph Weizenbaum's computer therapist ELIZA, the project repeats back what it thinks it has heard from the user, leading to surreal, uncanny and absurd conversations.
More information on this project here.