AI for Good stories

Creating community-driven datasets: Insights from Mozilla Common Voice activities in East Africa

The Mozilla Foundation has been working with the GIZ FAIR Forward initiative to promote the creation and utilization of open voice data and technology in Kinyarwanda, Kiswahili and Luganda. These efforts include the crowdsourcing of large voice datasets together with local communities. The objective...

Featured Image

The Mozilla Foundation has been working with the GIZ FAIR Forward initiative to promote the creation and utilization of open voice data and technology in Kinyarwanda, Kiswahili and Luganda. These efforts include the crowdsourcing of large voice datasets together with local communities. The objective of this review is to document the strategies that have been deployed to create publicly available voice datasets in these three communities using the Mozilla Common Voice platform. It aims to provide existing and future voice communities as well as organizations who support them, with insights and recommendations, by exploring not only the necessary technical steps but also the social dynamics and structures at work.