Language-data infrastructure
A structured foundation for organizing language resources and their provenance.
GamVoiceGambian languages. Shared possibilities.
GamVoice is building language-data and natural-language-processing infrastructure for Gambian languages.

Our starting point
We are starting with four Gambian languages. Careful language work starts with the people who know and use them.
Meet our language focus →Why this matters
AI increasingly shapes how people communicate, learn, create, and access information. Generative systems make interaction easier through natural language, but access alone does not close the language-data gap.
Many African languages remain underrepresented in the digital resources and research behind these systems. Working from The Gambia, GamVoice is helping build the foundations for Gambian languages to participate more meaningfully in the AI ecosystem.
The work ahead
Useful language technology needs more than a model. We are developing the resources and workflows that make responsible research possible.
A structured foundation for organizing language resources and their provenance.
Developing an approach to speech resources with appropriate consent and clear purpose.
Working toward careful, consistent written representations of spoken language.
Building a domain for language labels, human review, and traceable decisions.
Exploring how language resources can support natural-language-processing research.
Planning documented resources with considered access, privacy, and release decisions.
A connected approach
Our intended development process keeps language expertise and human judgment at the center.
Research is a shared effort
We are interested in working with researchers, language experts, educational institutions, and responsible technology collaborators. Local context and linguistic knowledge should help shape the questions we ask.
Discover our research interests and the kinds of collaboration we hope to develop.
Explore collaboration →Our development principles
Appropriate consent and clear purpose should guide data collection. Private material should stay private unless a release is explicitly approved.
Transcription and annotation need linguistic judgment, quality checks, and traceable review decisions.
Controlled access and documented provenance form part of the foundation. Collection and release practices still require careful review.
A foundation for what comes next
Explore the purpose behind GamVoice and our direction for responsible language research.