Closing Language Gap in Africa with AI Systems

I’m so pleased to see LINGUA expanding to Africa.   Too many African languages still have only a limited digital presence, which means their speakers risk being left behind as AI develops.  LINGUA is about helping strengthen local-language datasets so that AI better reflects the linguistic and cultural diversity of the societies it serves. This is essential for inclusion and opportunity.   I’m looking forward to the ideas that emerge from this call for projects. 

Across Africa, thousands of languages remain underrepresented in modern AI systems. When languages are not well represented in datasets, models, and tools, communities face barriers to accessing education, healthcare information, public services, financial inclusion, and other opportunities increasingly shaped by digital technologies. To help address this gap, Microsoft AI for Good Lab, Gates Foundation, the Masakhane African Languages Hub, and Google.org are joining forces to launch LINGUA Africa, an open call designed to strengthen the language foundations needed for inclusive AI in Africa. Applications close 15 June ✨

📢 MULTIDISIN Webinar Series (starting on 18th May 2026), a series of interdisciplinary conversations exploring how dis/misinformation operates across multilingual and culturally diverse contexts 📢 MULTIDISIN is a Research Ireland New Foundations-funded project led by Brendan Spillane and scholars at Dublin City University Iker Erdocia DCU SALIS and University College Dublin, and supported by the DCU FuJo Center and the ADAPT Centre. The project examines dis/misinformation from a comparative perspective, with a particular focus on medium-sized and minority languages (including Irish), bringing together expertise from linguistics, computer science, journalism and media studies. The webinar series features internationally recognised scholars and practitioners working on multilingual disinformation, generative AI, fact-checking, digital exclusion and platform governance. 📅 Starting on 18th May 2026 📍 Online (Zoom) 🌍 Open to all ✉️ Registration: iker.erdocia@dcu.ie Please see the programme for full details: https://lnkd.in/e9EdNvF3 #Disinformation #Multilingualism #DigitalMedia #AI #FactChecking #LanguagePolicy #MinorityLanguages #MediaStudies #DigitalGovernance #ResearchIreland #Linguistics

The first time I spoke to Sharon Ibejih, I couldn't stop taking notes. Many months later, we came back and asked her (and Tonative) to join us in bringing The Artificial Future to life. Since then, Tonative (Sharon Ibejih, Cynthia Amol & Cynthia Thuo) has hosted a workshop for our community, opening people up to a new world of data and linguistics in an AI context. One of the coolest parts of our Hackathon is the special Young People In Tech (YPIT) x Tonative storytelling track - gathering stories and essays in our African local languages to contribute to language data for low-resource languages. And to crown it all, Sharon is going to sitting on the Keynote panel at the #YPITAF conference on June 13th - talking about the role that data plays in building a functioning AI ecosystem. Talk about an all-round partner! We're glad to have you on Sharon! 🚀

I'm excited to be joining the panel at The Artificial Future Conference, hosted by Young People In Tech (YPIT). I'll be discussing "The State and Future of AI in Africa and Globally" alongside leaders working across data, infrastructure, applications and policy. At Tonative, my work focuses on African language dataset curation and community-driven approaches to improving language representation in AI. As AI adoption accelerates across the continent, one question remains critical: How do we continuously improve the representation and diversity of African languages in AI models? I said yes to this conversation because building AI for Africa requires more than models and compute. It requires the data, linguistic expertise and community participation needed to ensure our technologies reflect the people they are intended to serve. I'm looking forward to sharing lessons from our work at Tonative, discussing the opportunities and challenges in African language AI, and learning from others who are helping shape the continent's AI future. If you're interested in the future of AI in Africa, I'd love for you to join us. 🎟️ Get your conference tickets: https://lnkd.in/eu5-BpdP #YPITAF #ArtificialFuture #AIAfrica #TechAfrica #LagosAI #BuildAfrica

Honoured to see Tonative community featured by the Deep Learning Indaba. ✨ This recognition reflects the work being done by native speakers, data validators, researchers, engineers, and all other contributors committed to building AI systems that truly represent African languages and cultures. From African language dataset curation to multilingual AI tools, we remain focused on creating technology that is inclusive, authentic, and community-driven. The future of AI in Africa must be built with African voices at the center and we’re proud to be contributing to that future. Thank you to the Deep Learning Indaba team for the recognition. 🌍 #DLI2026 #DeepLearningIndaba #AIForAfrica #AfricanLanguages #LanguageTechnology #DigitalInclusion #Tonative

Community Showcase: Tonative ✨ What happens when native speakers, researchers, engineers, and linguists come together to shape the future of African AI? You get Tonative. From contributing African languages to global benchmarks like Global PIQA and XNLI, to building tools like VOLANG (a real-time speech-to-speech translation tool), they are ensuring African languages are represented accurately, authentically, and inclusively in modern AI systems. Their community-driven approach puts native speakers at the centre of dataset validation and evaluation, creating AI that truly sounds like us. Beyond research, their contributions at NeurIPS, CHI, and the Deep Learning Indaba are advancing language technology across the continent. Watch their community showcase to see how they are building sovereign AI for African languages and cultures. Join #DLI2026 In-Person Registration: Still open through the Ticket Lottery (paid) until 30 May 👉 https://lnkd.in/dP8ZD7fB Virtual Indaba: Registrations are open to join #DLI2026 online! 👉 https://lnkd.in/d2X2pVbG #DLI2026 #Indaba2026

"$250,000 for African Language AI Projects LINGUA Africa funds projects that build the open data, models, and applications African languages need to participate in modern AI. The program goal is reaching one billion Africans with locally relevant AI tools by 2029. Three grant categories are open: Data creation: building, curating, translating, or licensing language datasets. Up to $50,000 in cash and $50,000 in compute credits. Model or tool development: creating or adapting models, benchmarks, and infrastructure for African languages. Up to $100,000 in cash and $100,000 in compute credits. Sectoral applications: piloting language technologies in real-world settings with a credible path to measurable impact. Up to $250,000 in cash and $400,000 in compute credits. Priority sectors include agriculture and food security, education, healthcare and public health, financial inclusion, and government and civic services. All supported projects must contribute openly licensed resources. Who Can Apply Eligible applicants include nonprofits, universities, research institutes, social enterprises, cultural organizations, startups, and consortia working in the public interest. Organizations based outside Africa can apply if they show meaningful partnership with Africa-based institutions or communities. Beyond cash, selected projects receive Azure and Google Cloud Platform credits and in-kind technical collaboration from Microsoft AI for Good Lab. Apply Now: Deadline is June 15, 2026 More Funding Opportunities Please sign up now to get our email updates. Learn how to get startup funding for your technology business, and find new funding opportunities with donors..." https://lnkd.in/dbxjyRhZ #metaglossia #metaglossia_mundus

Every AI pitch deck for Africa assumes the models speak African languages. They don't. Every mockup in Kinyarwanda I've built the first testers say the same thing: it doesn't make sense. It sounds off. The Kinyarwanda's bad. The largest foundation models handle a handful of the continent's 2,000 languages with any fluency. Swahili, barely. Hausa, Yoruba, Amharic — thin to useless. The "AI for Africa" demo runs in English. The investors watching don't speak the local languages — so they can't tell the words are wrong. We're building African AI in languages that aren't native to Africa. The scale opportunity doesn't exist until the language gets solved.

A new study shows that well-tuned Small Language Models (SLMs) can outperform prompted large language models on citation verification tasks across multiple languages. Even more striking, models trained primarily on English data were able to generalize effectively to lower-resource languages with minimal. In "Multilingual and Cross-Lingual Citation Needed Detection on Wikipedia for Lower-Resource Languages," the researchers demonstrate that capability is not always a function of scale. For many multilingual and low-resource use cases, smaller, specialized models can deliver stronger performance while remaining more affordable and accessible to deploy. For organizations building AI in resource-constrained environments, this is an important reminder: the future of AI may depend less on building bigger models and more on building the right ones. Read the full paper >> https://lnkd.in/ed-KYYAr Gerrit Quaremba Amy Rechkemmer Elizabeth Black Denny Vrandečić Elena Simperl King's College London Wikimedia Foundation Association of Language Companies in Africa Languages Africa Africa Languages and Beyond #SmallLanguageModels #SLMs #LanguageAI #LowResourceLanguages #MultilingualAI #AIResearch #DigitalInclusion #SovereignAI #EqualyzAI