[multi’vocal] works towards an alternative future for digital voice-assistants by creating performative synthetic voices that can present with more than one age, gender and geographic origin.
Already we have collected 3,500+ voice recordings from people in Denmark and through our website, and we want to continue making space for people who are underrepresented in voice design, especially marginalised and underrepresented communities. To do this, during the fellowship, we will investigate new formats for developing and presenting synthetic voices that can reach beyond the tech-community, and investigate different machine learning methodologies for generating different sonic presentations of [multi’vocal] voices. We use our recording sessions to catalyse and capture participants’ reflections on the aesthetics and politics of synthetic voices on their lives, and we aim to present this research as an installation and release a vinyl record for broader dissemination.
Living in a world where machines talk to us with synthetic voices, we believe it is important to consider and discuss representation and aesthetics. Most of the voices on our devices and in our systems are designed to have binary vocal identities. It doesn’t have to be this way. Our project reimagines synthesised speech by presenting diverse and collective voices, thereby stimulating reflection on the sonic appearances of future synthesised speech that goes beyond the binary.
We are an interdisciplinary activist collective with a feminist agenda based in the humanities, computer science, and sound and interaction design. The rapidly growing field of voice-based interactions with technology is largely controlled by a few commercial tech giants, making participation in voice design difficult and undemocratic. We believe it is highly important to challenge the status quo, investigate alternatives that go beyond the binary, and spark debate.