AI-generated voices are increasingly present in both everyday life and the arts, surpassing the polarisation of natural versus artificial. While remarkable, the technology's development and artistic use focus mainly on human imitation and assistive tools, frequently falling short of exploring its inherent potential. Therefore, this thesis examines the practical and theoretical dimensions of artificial voice-sound phenomena, focusing on techno-cognitive engagement in forming hybrid voice identities as an ongoing process of becoming.

Central to this study is a critical examination of the shift from unpredictable, individual voices to predictable, non-individual ones. Through performative works and installations, using various technologies including an IKOsahedral loudspeaker—a unique, 20-sided loudspeaker instrument that utilises beamforming algorithms (είκοσι, íkosi = 20)—I expand artistic methods and insights in producing poetic artifacts, virtual voices, and augmented realities.

These works are interwoven with Vilém Flusser’s notion of techno-imagination, Michel Chion’s theory of the audiovisual contract, Michel de Certeau’s concept of vocal utopias, Mikhail Bakhtin’s heteroglossalia, and Gilbert Simondon’s philosophy of individuation, synthesising both intellectual and artistic frameworks into a novel theoretical paradigm. This paradigm allows for questioning identity formations in generative ways, navigating the interstitial space between the private and public, the human and the technological.

Furthermore, my research seeks to conceptualise the instrumentality and multiplicity of artificial voice by reimagining existing sonic and aesthetic terminologies and applying the resulting interlocked concepts to my practice. In addition, I propose and analyse a form of ‘virtuosity’ linked to Paolo Virno’s philosophy of the performative-virtuosic within post-Fordist labour,
aiming to develop a more differentiated vocabulary for addressing the socio-cultural implications of AI. This includes a critical examination of the risks associated with the commercialisation and standardisation of voice, given that AI development is presently dominated by a small elite within major tech companies.

I expect the results to contribute to the field of performance practice, sonic theory, voice studies, and discourses on how voice performs in mediatised environments.