Multiple Selves is an experimental sound piece that contains three songs titled “Glitch Ghosts,” “Glitch is Error,” and “Ghost Echoes.” These songs were generated by a text-to-speech application created by the artist and three choir members mimicking the sound of an artificially generated voice. This project can be presented as a three-channel video installation or as a live performance by the choir.

The text-to-speech application is based on the International Phonetic Alphabet of English words. Unlike other computer voices or commercial text-to-speech applications, the syllables were exaggerated by randomly alternating pitch, length and tone of the pronunciations. Songs were generated by the artist typing text of her own and quotations from the book Glitch Feminism by Legacy Russell.

The possibility of gender-ambiguous computer-generated voices has not been widely considered because of the commercial nature of their use. Rather, Siri or other similar AI voices are more likely to reinforce gender prejudice. When we communicate with virtual assistants, we command, and they answer. Even though we know this is not a real human, we still engage in a conversation as if speaking to a human being because this technology triggers the language function in our brain the same way talking to a real person would do (Nass & Brave 2007).

This project aims to create computer-generated voices that suggest the possibility of non-gendered voices. The concept of “Multiple Selves” is taken from the book Glitch Feminism but is originally found in E. Jane’s Nope (a manifesto). “Multiple Selves” “pushes back against a flattened reading of historically othered bodies—intersectional bodies who have travelled restlessly, gloriously, through narrow spaces.” Multiplicity acts like freedom. Instead of making a singular computer voice, this project embraces the concept of self as multiplicity and approaches voice by differing all the pitches and tones of single phonetics. As such, instead of making computer-generated voices that mimic humans, humans (Phth ensemble) learn and mimic computer-generated voices.