Just Because We Camp, Doesn't Mean We Should: The Ethics of Modelling Queer Voices

TL;DR

This paper investigates the ability of voice cloning models to accurately synthesize queer voices, revealing homogenization effects and discussing ethical concerns related to data collection and potential harms.

Contribution

It demonstrates the limitations of current voice cloning models in capturing queer voice styles and initiates a discussion on the ethical implications of modeling marginalized voices.

Findings

Synthesized speech sounds less 'gay' than original for LGBTQ+ listeners.

Loss of 'gay voice' reduces perceived similarity to original speakers.

Homogenization effect impacts accessibility and authenticity.

Abstract

Modern voice cloning models claim to be able to capture a diverse range of voices. We test the ability of a typical pipeline to capture the style known colloquially as "gay voice" and notice a homogenisation effect: synthesised speech is rated as sounding significantly "less gay" (by LGBTQ+ participants) than its corresponding ground-truth for speakers with "gay voice", but ratings actually increase for control speakers. Loss of "gay voice" has implications for accessibility. We also find that for speakers with "gay voice", loss of "gay voice" corresponds to lower similarity ratings. However, we caution that improving the ability of such models to synthesise ``gay voice'' comes with a great number of risks. We use this pipeline as a starting point for a discussion on the ethics of modelling queer voices more broadly. Collecting "clean" queer data has safety and fairness ramifications, and the resulting technology may cause harms from mockery to death.