Participatory speech dataset creation methods improve queer community inclusion
Towards participatory speech dataset curation: A queer case study and conceptual framework
Artificial IntelligenceComputers and Society
Summary
Collecting voice data often misses the needs and concerns of queer people, especially due to past harmful AI uses. The authors look at current speech data methods and why they may not work well for the LGBTQIA+ community. They review efforts to include queer voices in AI and speech data work. From this, they suggest a new way to create speech datasets that involves queer and marginalized communities as active participants who help shape the project and keep control over their data.
What this means in practice
- •For voice interface developers: Design speech recognition systems using datasets created with queer community participation to reduce bias and increase relevance for diverse users.
- •For diversity and inclusion teams: Create data collection policies that empower marginalized communities to co-shape AI projects, helping ensure respectful representation and reduce harms.
A position paper. It proposes an approach and reports no results.
Authors
Brooklyn Sheppard, Anaelia Ovalle, Adina Williams, Levent Sagun
Abstract
In this paper, we motivate the need for a participatory speech dataset creation framework through a case study of the LGBTQIA+, or queer, community - a community with documented concerns about AI and reported harms, including attempts to develop 'gaydar' technologies that purportedly identify individuals as queer. We review common speech data collection practices, why these methods may be unsuitable for engaging with queer speakers, and discuss previous efforts in participatory AI with queer community engagement, as well as participatory endeavours specific to speech data collection for other marginalized communities. From this review, we develop a conceptual framework for participatory speech data curation by, for, and with marginalized communities drawing on insights from co-design and knowledge sharing. We propose a framework comprising overlapping and two-way processes of defining a community, project formulation, modes of participation, and personal autonomy.