Artificial Intelligence and Voice

AI personal assistants have been almost exclusively assigned female voices. For example, Siri, Alexa, Cortana, and Echo. Yet, their vocal utterances were not aiming to be exactly human-like, deliberately allowing for glitches in intonation and pitch that give away their machinic quality. This continues a tradition in the use of female voices assigned to serving machines across the history of technology and communication, raising important questions about gender and feminism in relation to automation. With the development of new digital assistants and their ability to switch between female and male voices, as well as the perfecting of their fluency and human-ness in their articulations and interactions with humans, the voice and technology debate could be said to stretch from feminism and gender specificity, on the one side, to questions to do with voice-based human-machine interrelations and their implications.

As AI devices develop human-like vocal accuracy, a new set of cultural, social, ethical, artistic and philosophical questions and implications begin to emerge: How such a technology might confuse and undermine voice-based verification systems? In an era of fake news, how does it complicate further existing cultural ideas about sources of information. How will users interact with the technology and what new creative processes, artistic responses and forms of sociability will develop from this interaction? How much will users trust it and what kind of new fears, paranoias will it inflict? This proposed project will research innovative methods by which the arts and humanities might address contemporary cultural processes about new forms of interaction between humans and machines as AI technology enters a new stage of vocal indistinguishability between them.

Supported by the TECHNE National Productivity Investment Fund (NPIF) AHRC Scholarship the Royal College of Art in partnership with IBM's iX lab is looking to award a PhD studentship to explore these issues of new communication interactions drawing on a methodology that relies on a combination of art, design and computer science approaches. In this research by practice, coding and predictive practices underpinning all voice systems will be subject to critical enquiry.