Community-Based Language Initiatives

#AiDays2025 Round Table : #Community Sourcing for low resource languages In an era where AI is fast shaping the contours of our digital future, VISWAM.AI initiative stands as a timely and transformational one. Their mission to build community-sourced Large Language Models (LLMs), grounded in India’s rich linguistic and cultural diversity, is not just pioneering—it’s redefining how inclusive and ethical AI should be built. By anchoring their work in community participation, linguistic preservation, and ethical co-creation, Viswam.ai offers a people-first approach to AI—moving beyond data extraction to cultural stewardship. Their ambition to mobilize 1 lakh community interns to collect data from underrepresented geographies across India is both bold and brilliant. This isn’t just about building better AI—it’s about building equity, agency, and cultural resilience through AI. 1. Linguistic Equity by Design In India, where linguistic hegemony often privileges English and Hindi, AI systems risk reinforcing this imbalance. The solution? Intentional design. Allocate equal engineering and validation efforts to low-resource languages. Ethical AI must be built on informed consent, community ownership, and fair compensation—because data is not just input, it’s identity and heritage. 2. Decentralized Internship Model By decentralizing AI development, we bridge the urban-rural digital divide. This model should focus on: Capacity building through training in ethics and digital literacy Inclusivity by involving women, Dalit and Adivasi youth Localized platforms using mobile-first tools in native languages Partnerships with Swecha, local NGOs, and institutions serve as trust bridges to ensure mentorship and sustainability. 3. Tools for Low-Resource Languages Many Indian languages are oral-first, with complex dialects and sparse corpora. Community-driven solutions—like collecting voice datasets from folklore, and crowdsourcing annotation—are key. Elders, poets, and storytellers become linguistic technologists, preserving not just language but legacy. 4. Trust & Transparency Bias in AI is structural. To mitigate it: Include diverse dialects and accents in training Conduct bias testing and community validation Promote explainable AI with local language dashboards and storytelling What’s Next? A living white paper on ethics, governance, and technical guidelines A roadmap for the internship program, with toolkits and impact metrics Collaboration with literary and linguistic organizations to enrich model depth VISWAM.AI is planting seeds for an AI movement rooted in language justice, data sovereignty, and community wisdom. Let’s co-create systems that don’t just understand our languages—but respect our voices. DC* Chaitanya Chokkareddy Kiran Chandra Ramesh Loganathan Centific

Can #AI Help Revitalize #Indigenous #Languages?, by Serena Jampel -- Indigenous technologists are using artificial intelligence to create tools that support language learning in ways that center community control and cultural values. Instead of replacing traditional teaching, these tools aim to reinforce everyday use and make learning more accessible, especially for younger people and schools with limited resources. The article highlights three examples developed within Indigenous communities: – A talking robot for Anishinaabemowin learners Danielle Boyer, an Anishinaabe roboticist, designed Skobot, a small robot that sits on the shoulder and speaks Anishinaabemowin. It uses motion and sound sensors to play pre-recorded words and phrases from fluent speakers. Boyer’s goal is to help children hear and interact with their heritage language in daily life, guided by cultural protocols and community input. – A Choctaw-English chatbot built for practice and humor Jacqueline Brixey, who is Choctaw and Cherokee, created Masheli, a conversational chatbot that talks in both Choctaw and English. It uses culturally familiar language, humor, and storytelling to help learners feel comfortable while practicing. The bot avoids correction and encourages continued engagement, especially for users who may be shy about speaking aloud. – Offline voice-based learning kits for schools Michael Running Wolf leads the FLAIR project in Quebec, which develops voice-activated curricula for Indigenous languages. These portable kits are designed for classrooms without reliable internet access. Using limited recordings from elders, the systems provide automatic speech recognition to guide pronunciation and support language instruction directly in community schools. All three developers emphasize that their AI tools are made in collaboration with elders and are shaped by local priorities, not commercial platforms. For more information: https://lnkd.in/d-64Epte 

Our paper describing the first 2 years of our #IndigenousAI project is now available (https://lnkd.in/dE6ThQ_z), describing a proposal of an AI development cycle specific for the context, our engagements with Indigenous communities in Brazil, and related technical discussions. Since 2022 we have been exploring application areas and technologies in which Artificial Intelligence (AI) and modern Natural Language Processing (NLP), such as Large Language Models (LLMs), can be employed to foster the usage and facilitate the documentation of Indigenous languages which are in danger of disappearing. We start by discussing the decreasing diversity of languages in the world and how working with Indigenous languages poses unique ethical challenges for AI and NLP. To address those challenges, we propose an alternative development AI cycle based on community engagement and usage. Then, we report encouraging results in the development of high-quality machine learning translators for Indigenous languages by fine-tuning state-of-the-art (SOTA) translators with tiny amounts of data and discuss how to avoid some common pitfalls in the process. We also present prototypes we have built in projects done in 2023 and 2024 with Indigenous communities in Brazil, aimed at facilitating writing, and discuss the development of Indigenous Language Models (ILMs) as a replicable and scalable way to create spell-checkers, next-word predictors, and similar tools. Finally, we discuss how we envision a future for language documentation where dying languages are preserved as interactive language models. This is joint work of IBM Research and the USP - Universidade de São Paulo through the C4AI - Center for A.I. # USP+IBM+FAPESP Brazil, by Claudio Pinhanez, Paulo Cavalin, Luciana Storto, Thomas Finbow, Alexander Cobbinah, Julio Nogima, Marisa Affonso Vasconcelos, Ph.D., Pedro Henrique Domingues, Priscila Mizukami, Nicole Grell Macias Dalmiglio, Majoi Gongora, and Isabel Gonçalves. Some info about the project: https://lnkd.in/d7bg4jPB #IndigenousAI #IBMResearch #ResponsibleAI #AI4Good #C4AI #USP Paper link: https://lnkd.in/dE6ThQ_z

India has over 1,600 languages and dialects, yet most AI systems only cater to the widely spoken ones. Now, startups like TuluAI, Aakhor AI, and KashmiriGPT are building AI tools for low resource languages from scratch. These founders are collecting data through storytelling sessions in rural areas, WhatsApp voice note drives, and community workshops. They're recording elders, teachers, and homemakers, capturing not just words but idioms, humor, and cultural context that major translation tools miss. The challenge is immense: Tulu's ancient script lacks Unicode standards. Kashmiri faces the risk of digital extinction. Bodo and Assamese have minimal online presence. This is more than tech innovation, it's cultural preservation through AI. When communities see their voices helping preserve their language, they feel ownership in the process. As India becomes ChatGPT's largest market outside the U.S., these startups remind us that true AI advancement isn't just about scale, it's about inclusion, representation, and ensuring no language gets left behind in the digital age. #AI #Linguistics #India #TechForGood #Innovation #CulturalPreservation https://lnkd.in/g3rbK2B7

NEW WAY TO KEEP OUR LANGUAGES ALIVE. Not in a museum. Not locked in a classroom. Not waiting for permission. In your pocket. The Coeur d’Alene Tribe just launched something powerful. An app called Coyote Stories. Built by their own people. For their own people. And here’s what I respect most… they didn’t wait for funding, a big team, or some outside “expert” to validate it. Two people built it. That’s it. Inside the app is real language, real voices, real stories. Flashcards, audio from speakers, stories told the way they’re meant to be told. Even Coyote himself guiding the experience. This isn’t just tech. This is survival. Because when a language is down to a handful of speakers, you don’t have time to debate strategy for ten years. You move. You build. You protect what’s left. And they understood something a lot of people still don’t. If our kids are on phones, then our language needs to be there too. Not watered down. Not turned into something it’s not. But carried forward in a way they’ll actually use. That’s how you keep something alive. They also made a choice that matters. No generic AI visuals. No stock imagery. They brought in their own artists, their own people, their own style. That’s the difference between content and culture. One is made to fill space. The other carries responsibility. And here’s the bigger picture. This isn’t just about one Nation. They’re already working with other Nations to help them build their own versions. That’s how this spreads. Not through institutions. Through community. Through example. This is what language revitalization actually looks like on the ground. Not reports. Not panels. Action. Build it. Record it. Share it. Teach it. Before it’s gone... 👇 Link to Coyote Stories app in comments 👇 #IndigenousInnovation #LanguageRevitalization #Storytelling #IndigenousTech #EaglespeakerPublishing

𝗚𝗹𝗼𝗯𝗮𝗹 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗿𝗲𝘃𝗶𝘁𝗮𝗹𝗶𝘀𝗮𝘁𝗶𝗼𝗻 𝗸𝗲𝗲𝗽𝘀 𝗳𝗮𝗶𝗹𝗶𝗻𝗴 𝗳𝗼𝗿 𝗼𝗻𝗲 𝘀𝗶𝗺𝗽𝗹𝗲 𝗿𝗲𝗮𝘀𝗼𝗻: 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗹𝗶𝘁𝗲𝗿𝗮𝗰𝘆, 𝗻𝗼𝘁𝗵𝗶𝗻𝗴 𝗶𝘀 𝘀𝘂𝘀𝘁𝗮𝗶𝗻𝗮𝗯𝗹𝗲 𝗮𝗻𝗱 𝗿𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗵𝗮𝘀 𝗯𝗲𝗲𝗻 𝘁𝗲𝗹𝗹𝗶𝗻𝗴 𝘂𝘀 𝘁𝗵𝗶𝘀 𝗳𝗼𝗿 𝗱𝗲𝗰𝗮𝗱𝗲𝘀. -Documentation alone does not save languages. -Orthographies alone do not save languages. -Emotion alone does not save languages. -Technology alone does not save languages. 🇳🇿 Māori (New Zealand) Revitalisation only stabilised once Kōhanga Reo (language nests) and Kura Kaupapa Māori integrated literacy-rich environments, producing fluent and literate generations. 🇨🇦 Inuit and First Nations (Canada) Regions that implemented Inuktitut literacy programmes achieved higher transmission rates than those focusing solely on oral revitalisation. 🇧🇴 Aymara & Quechua (Bolivia) Education reforms integrating mother-tongue literacy led to increased school retention and a rise in community-led writing and publishing. 🇳🇴 Sámi (Norway/Finland/Sweden) Media, literature and digital tools became impactful only after widespread Sámi literacy training. 🇧🇷 Indigenous Amazonian communities Projects that linked literacy with agricultural knowledge, governance and health practices saw stronger linguistic resilience. what actually works? 1️⃣ Community-based literacy programmes Research shows that literacy succeeds when communities lead instructional design and content creation (Hinton, 2013). 2️⃣ Early childhood immersion + literacy exposure The Māori, Sámi and Hawaiian models all prove that literacy must begin early to anchor long-term revitalisation. 3️⃣ Culturally grounded reading materials Books and digital content must reflect local realities, not imported or generic themes (Grenoble & Whaley, 2006). 4️⃣ Literacy linked to livelihood When reading/writing the language helps with agriculture, land rights, market access or local governance, adoption increases dramatically. 5️⃣ Integration of ICTs and AI Tools supporting reading in minority languages (keyboards, spell checkers, audio-text apps, OCR for local scripts, AI-readers) improve daily use and visibility. 6️⃣ Training local educators, not only linguists Hornberger emphasises that literacy becomes sustainable only when local teachers, elders and youth facilitators are empowered. 👉🏾 A minority language without literacy remains dependent on outsiders. A minority language with literacy becomes self-sustaining. If we want languages to live, not merely to be preserved, literacy must return to the centre of global revitalisation strategies. #MinorityLanguages #IndigenousLanguages #LanguageRevitalisation #MotherTongueLiteracy #Biliteracy #Fishman #Hornberger #EndangeredLanguages #UNESCO #LanguagePolicy #CulturalRights #GlobalEducation #LinguisticJustice #CulturalDiversity #AIForLanguages #DigitalHumanities #SustainableDevelopment #EducationForAll #LanguagePreservation #CommunityEmpowerment

For the 1.2 billion people who communicate using low-resource languages, #AI applications are often not locally relevant and in many cases, not even accessible.    A few days ago, Google rolled out 110 new languages to Google Translate, including a number of African and Indian local languages with the help of #AI (see more here: https://lnkd.in/e9aTxYzr ). And they are not alone. Masakhane, Lelapa AI., GhanaNLP, Indigenous in AI/ML,OpenAI, Lacuna Fund, Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH FairForward and others are already contributing to making low-resource language data more widely available.    Even so, the opportunity to preserve and promote local languages by integrating them into the rapidly advancing #digital landscape remains largely unexplored.    At UNDP, we are working to make sure that no community is left behind in the #AI revolution. By collaborating with players across local ecosystems as equal partners, we can help preserve local languages and shape the development of AI to benefit everyone, everywhere.   Stay tuned and share in the comments if you know of similar or complementary efforts.

Taking Santhali Digital: What It Teaches Us About Language, Technology, and Learning A remarkable story is unfolding in the heart of eastern India, where coders, teachers, and grassroots educators are scripting a new future for one of India's oldest languages; Santhali. Having led a digital learning initiative in Jharkhand, I have seen firsthand how language sits at the heart of all meaningful education. It is not just a medium of instruction. It is a vessel of worldview, memory, and belonging. Yet it took over a hundred years since Ol Chiki’s creation for the script to find real traction in the digital world. Why? Because our digital infrastructures, Unicode standards, keyboards, operating systems, have been built without the multilingual, oral, and script-diverse realities of India in mind. What this movement around Santhali teaches us is clear: (i) Learning in one’s mother tongue builds not just comprehension, but confidence (ii) Community-led digital innovation is often more sustainable than top-down (iii) Edtech must support scripts like Ol Chiki to be truly inclusive interventions The future of education in India lies not in scaling uniformity, but in honoring multiplicity. When every child sees their language on a screen, they know they belong in the classroom and in the world. #Santhali #OlChiki #EdTech #LanguageJustice #Jharkhand #MultilingualEducation #DigitalInclusion #GrassrootsInnovation

Language is more than communication. It is identity, culture, and heritage. India has over 700 tribal communities, speaking 461 languages and 71 distinct mother tongues. Among these, 82 are vulnerable and 42 are critically endangered. IIIT Hyderabad has taken a pioneering step with the Adi Vaani project, India’s first AI-powered translator for tribal languages. The beta version supports Santali, Mundari, Bhili, and Gondi. Native speakers helped refine the translations to ensure cultural authenticity. The project includes text-to-speech tools and translation systems between English, Hindi, and these tribal languages. It is designed to make educational, healthcare, and government resources accessible in low-resource languages. What excites me is how AI can go beyond enterprise or tech use cases. Here, it is preserving culture, enabling inclusion, and creating real impact at the grassroots level. . . . #AIforGood #DigitalInclusion #CulturalPreservation #TribalLanguages #AI #EdTech #Innovation #SocialImpact

When Communities Lead, Change Lasts For years, outsiders led the change. Then one village said... “Not anymore.” For years, Fatou’s village in Tambacounda, Senegal, had hosted visitors with clipboards and promises. Projects came. Projects went. Each began with good intentions and ended the moment the funding did. So when a new initiative arrived, the villagers were polite but skeptical. This one felt different. There were no long speeches. No handouts. Just a question: “What kind of future do you want to build?” That question changed everything. Through the Community Empowerment Program, held in Wolof and Pulaar, villagers began meeting twice a week beneath the acacia tree. At first, they discussed reading, writing, and basic health. Then they moved on to larger ideas, such as human rights, governance, and enterprise. They elected a Community Management Committee consisting of men and women trained to lead and sustain their own initiatives. And slowly, something shifted. Girls who once dropped out of school returned. Women pooled savings to start new markets. Men began to speak differently about their daughters’ futures. To coordinate across villages, committees used local radio and simple mobile phones to share lessons, updates, and collective goals. It wasn’t flashy technology, but it worked. In a matter of months, a cluster of villages had transformed into a self-organized learning network. No consultants. No dashboards. Just connection, powered by trust. By the program’s end, independent studies found: - 121 communities across Senegal had joined - 60% lower rates of domestic violence - 40% more women's participation in leadership When Fatou stood before her neighbors to close the final session, she smiled and said, “Before, people came to teach us. Now, we teach each other.” And in that moment, the ownership was unmistakable. The Lessons? 1️⃣ When communities design in their own language, set their own pace, and utilize simple technology to stay connected, support becomes an amplifier. 2️⃣ When data begins to flow from the ground up, through radio messages, meeting logs, or WhatsApp updates, measurement becomes meaning. 3️⃣ When change belongs to the people who live it, it doesn’t fade when the funding ends. Collaboration builds momentum. Ownership makes it last. Local ownership turns impact projects into movements, and movements into lasting systems. #ScalingImpact #LocalOwnership #CommunityEmpowerment #TechForGood #AfricaInnovation #SocialImpact