Free the Voices: the first bank of diverse synthetic voices
An initiative that uses AI to create synthetic voices representing the diversity of the LGBTQ+ community, promoting inclusion and reducing bias in voice technology.
Voice-based discrimination is still a reality that especially affects the LGBTQ+ community. Voices that don’t fit prevailing social norms often face prejudice and microaggressions, leading many people to change or hide their true voice. In a world where voice technology is increasingly present, it’s crucial that synthetic voices reflect the true diversity of our society.
That’s why the “Free the Voices” project, a collaboration between LLYC and Monoceros Labs with its product Fonos, set out to create a bank of diverse TTS voices to:
Normalize voice diversity in digital environments.
Reduce bias through greater exposure to diverse voices.
Create more inclusive spaces in technology.
The problem: voice discrimination and lack of representation
The voice is one of the traits that generates the most bias and discrimination against the LGBTQ+ community. Some studies show that:
41% of gay men and 6% of lesbian women say they’re identified as LGBTQ+ because of their voices.
People whose voices don’t match prevailing gender expectations are often perceived as less competent or less serious.
This discrimination has an impact in many areas:
Professional life.
Public life.
Psychological and emotional wellbeing.
The solution: creating diverse voices with AI
“Free the Voices” is an initiative developed with LLYC to create a bank of diverse synthetic voices. The project collected more than 500 voices from the LGBTQ+ community in 12 countries to generate synthetic voices that:
Don’t identify any specific person.
Represent natural vocal diversity.
Can be used in many digital contexts.
This bank of diverse synthetic voices, available in Fonos, is made up of the following voices: Libertas and Vega, designed exclusively for LLYC, which together with Cástor, Fulu and Hatysa make up the first five diverse TTS voices in Spanish.
Impact and reach
The synthetic voices created are available in Fonos for use in:
Virtual assistants.
Audiovisual content.
Voice-overs.
Digital experiences.
Voice technology helps us break down
barriers in communication with technology
and with people, and that’s also why it
helps us break stereotypes. That’s why
building AI-based voice models to create
voices that are diverse in gender and accent
is crucial to building a future in which
technology and digital content are
more inclusive.
This project is a significant step toward:
Normalizing voice diversity in technology.
Reducing bias and discrimination.
Creating more representative digital spaces.
Empowering traditionally marginalized voices.
Technology and innovation
The project uses Monoceros Labs’ own generative AI technology, which lets us generate new, custom voice identities from diverse voice samples.
Main partners
LLYC: strategy, creative direction and project management.
Monoceros Labs: technology development of the speech synthesis models, and design of two new diverse identities: Vega and Libertas.
LGBTQ+ organizations such as Redi (Spain), Pride Connection México, Casa Frida (Mexico) and the Cámara de la Diversidad de Colombia: advice and community participation.
Awards and recognition
[El Mundo] The 100 Best Ideas of the Year: “17. The first bank of LGBTQ+ synthetic voices.” El Mundo website (in Spanish).
