AI built on local knowledge and shaped by community priorities.
About the Localized AI Studio
The Localized AI Studio is an open platform where communities, technologists, and practitioners come together to build, adapt, and deploy AI tools rooted in local knowledge, languages, and priorities. The Studio brings together three distinct but connected offerings, each designed to ensure AI serves the people who use it, on their own terms.
Through the Localized AI Studio, a diverse community of technologists, community practitioners, and civic actors will build and adapt AI tools that reflect local realities and strengthen the agency of the communities that use them.
Why it exists
Most AI tools are built far from the communities they are meant to serve, trained on data that does not reflect local languages, cultures, or realities. The result is technology that often excludes, misrepresents, or simply does not work for the people who need it most.
There is an urgent need to build AI from the ground up, rooted in local knowledge, shaped by community priorities, and designed to strengthen agency rather than undermine it.
What we offer
01
Community AI Lab
FreeOnlineCollaborative
Build AI that reflects your community
A collaborative space where communities and technologists work together to build, adapt, and test AI tools rooted in local languages, knowledge, and priorities. Participants gain hands on experience with AI development, grounded in their own context, and leave with tools they own and can sustain independently.
Get involved02
Language and Data Sprints
In-personIntensiveCommunity-based
Bringing AI development into your community
Short, focused sprints run in communities to collect local language data, test AI tools, and build the technical capacity needed to sustain locally grounded AI development. Participants leave with a stronger understanding of how AI works, how it can work for them, and the agency to shape how it develops in their context.
03
AI Adaptation Programme
CollaborativeContext-drivenLong-term
From existing tools to locally owned solutions
Organisations and communities identify an existing AI tool that could serve their needs, and work with our team to adapt, localise, and deploy it in their context. The focus is on co-creation, local ownership, and building the technical confidence to maintain and evolve the tool independently.
How the programme works
01
Identify
A community or organisation identifies an existing AI tool with potential for local use.
02
Assess
We assess the tool together, looking at language, data, context, and values alignment.
03
Adapt
We work alongside the community to adapt, localise, and test the tool in their context.
04
Deploy and own
The community deploys the tool and builds the capacity to maintain and evolve it independently.
What you take with you
Practical skills in AI development, adaptation, and deployment
Locally grounded tools built or adapted for your context
A peer network of technologists and practitioners
The technical confidence to sustain and evolve AI tools independently
The agency to shape how AI develops in your community
What we are working toward
A growing library of locally grounded AI tools, openly shared
Communities with the agency to build, adapt, and own the AI tools they use
Local language data sets that reflect the diversity of communities we work with
Evidence and learnings from each programme shared openly with the broader field
AI development shaped by the communities it is meant to serve
Who is it for
Community organisations, technologists, civic actors, and practitioners looking to build or adapt AI tools rooted in local knowledge and priorities. The Studio is especially suited to communities and organisations currently excluded from AI development, who want to shape how AI works in their context, on their own terms.
AI built on local knowledge, open to all
Every dataset, model, and tool we build is made available so communities and practitioners can keep building on their own terms, without dependency.
2
Languages modeled
2
Open-source datasets published
60
Hours of voice data contributed
2
Tools adapted and deployed
Also part of our work