Tech Giant’s Move Marks Milestone for Local-Language AI Accessibility Across the Continent
In a significant development for digital accessibility across Africa, Google has announced the expansion of its AI-powered search features to support 13 African languages. The rollout brings AI Overviews and AI Mode capabilities to languages including Kiswahili, Somali, Amharic, Hausa, and Yorùbá, enabling millions of users to interact with artificial intelligence tools in their native tongues.
The expansion marks a major step toward addressing what experts describe as a fundamental equity gap in AI access. For too long, artificial intelligence technologies have primarily served speakers of dominant international languages, particularly English, leaving hundreds of millions of Africans unable to fully participate in the digital revolution unfolding globally.
The newly supported languages include Kiswahili spoken in Kenya and Tanzania, Somali used in Somalia and Kenya, Afrikaans, Akan from Ghana, Amharic in Ethiopia, Hausa in Nigeria, Kinyarwanda in Rwanda, Afaan Oromoo in Ethiopia, Sesotho in Lesotho and South Africa, Setswana in Botswana and South Africa, Wolof in Senegal, Yorùbá in Nigeria, and isiZulu in South Africa. These languages were selected based on strong and growing search usage patterns across the continent, targeting active communities in countries including Kenya, Nigeria, Ethiopia, Tanzania, Ghana, Rwanda, Botswana, Senegal, and Somalia.
Two Tools, One Goal: Democratizing AI Access
The expansion encompasses two distinct features designed to make AI more accessible and useful for African users. AI Overviews provide concise, AI-generated summaries at the top of search results, offering quick answers while maintaining links to reliable sources for deeper exploration. This feature helps users quickly grasp complex topics without wading through multiple web pages.
AI Mode takes the experience further by enabling a more conversational interaction. Users can engage with the AI through text, voice, or even image sharing to receive detailed, personalized responses in their preferred local language. This interactive capability represents a significant advancement in making AI feel less like a distant technology and more like a practical tool for everyday problem-solving.
“When technology only speaks a dominant international language like English, it marginalizes millions of people whose first languages reflect a different culture, identity, and way of understanding information,” said Kabelo Makwane, Country Director for Google South Africa. “No one should be excluded from the AI economy because their first language isn’t English. When Africans can search, learn, and build in their own languages, AI becomes a driver of inclusive growth.”
Real-World Impact: From Classrooms to Market Stalls
The practical applications of this expansion are far-reaching. Students and teachers across Kenya can now ask complex questions and receive explanations in Kiswahili rather than struggling with English-language resources. Entrepreneurs in Somalia can explore market information and business insights using their native language. A farmer in Nigeria seeking advice on crop management can now interact with AI in Hausa, while a researcher in Ethiopia can conduct searches in Amharic.
For users across the continent, accessing these features is straightforward. The tools are integrated into the existing Google ecosystem, available through the Google mobile app or a web browser. Users simply tap on AI Mode within the search interface and type or speak their query in their preferred language. The AI system then generates responses and summaries tailored to their linguistic and cultural context.
Built on Solid Foundations: The Waxal Project
This expansion builds on years of foundational research through Google’s Waxal language project. The name Waxal, meaning “to speak” in Wolof, reflects the initiative’s core mission: making digital communication genuinely inclusive and representative of local cultures and languages.
Launched in February 2026, Waxal represents a groundbreaking open-source dataset providing more than 11,000 hours of speech data across 21 African languages. Developed over three years in partnership with leading African institutions including Makerere University in Uganda, the University of Ghana, Digital Umuganda in Rwanda, and the African Institute for Mathematical Sciences, the project combines machine learning research, linguistic expertise, and community collaboration.
What sets Waxal apart is its approach to data ownership and sovereignty. Unlike typical Big Tech initiatives where companies retain control over collected data, Waxal ensures that African partner institutions retain ownership of the data they collected. This framework represents a rare move toward digital sovereignty, enabling local institutions to become hubs of AI infrastructure rather than mere data suppliers.
The dataset includes approximately 1,250 hours of transcribed natural speech for automatic speech recognition, along with over 180 hours of high-quality recordings for text-to-speech synthesis. Rather than using scripted readings, researchers asked participants to describe visual stimuli in their native languages, capturing authentic linguistic variations including tonal nuances and code-switching patterns common in multilingual African communities.
Addressing a Critical Digital Divide
The significance of this initiative cannot be overstated. Sub-Saharan Africa is home to more than 2,000 distinct languages, yet only a handful have received the attention and resources necessary for natural language processing. This scarcity has created a profound digital divide, effectively excluding hundreds of millions of people from accessing voice-enabled technologies that have transformed how people in other regions interact with computers.
Speaking at the ITWeb Artificial Intelligence Summit last year, Neda Smith, Chartered Chief Information Officer of Agile Advisory Services, emphasized the devastating impact of this exclusion. The lack of African language inclusion in AI systems, Smith noted, has been shutting out millions of Africans from the digital world, highlighting the pressing need to bring African languages to the forefront of AI innovation.
Google’s announcement comes amid growing efforts across the tech industry to address linguistic representation on the continent. Microsoft recently introduced Paza, a pipeline and benchmarking tool for 39 African languages. Nigeria launched N-ATLAS in 2025, an open-source model capable of transcribing speech in Yoruba, Hausa, Igbo, and Nigerian English. Meanwhile, African startups are developing voice recognition and translation solutions aimed at local needs.
Technical Challenges and Community Solutions
Building AI systems for African languages presented significant technical challenges. Many African languages are linguistically rich with multiple layers of context, tonal variations, and complex grammatical structures that differ fundamentally from Indo-European languages that have dominated AI development.
Google’s approach centered on community collaboration rather than top-down implementation. African academic and community organizations led the data collection effort entirely, guided by Google experts on world-class data collection practices. This collaborative model ensured the corpus was built by and for the communities it serves, with each partner focusing on specific language subsets based on their local expertise.
The University of Ghana focused efforts on eight languages, while Makerere University collected automatic speech recognition and text-to-speech data for nine different languages. Digital Umuganda, in partnership with Addis Ababa University, led collection efforts for several major languages, while regional studios contributed high-quality voice recordings.
Looking Ahead: A Foundation for Innovation
Google has made clear this expansion represents just the beginning of a long-term commitment to building African-language AI capabilities that are both technically advanced and culturally grounded. The company is inviting users across the continent to try the new features and share feedback to help refine how AI systems understand and respond in their languages.
The Waxal dataset, now openly available on the Hugging Face platform under permissive licensing, provides a foundation for researchers, students, and entrepreneurs to build their own AI-powered applications. Local organizations are already leveraging the data for diverse use cases, from maternal healthcare research to educational tools designed specifically for African contexts.
For Joyce Nakatumba-Nabende, a professor and researcher at Makerere University, the implications are clear: “For AI to have a real impact in Africa, it must speak our languages and understand our context.” This principle now guides a growing movement to ensure that as artificial intelligence reshapes the global economy, African communities are not merely consumers of foreign technology but active participants in shaping AI’s development and deployment.
The expansion of Google’s AI search tools to African languages represents more than a product update. It signals a recognition that true technological progress must be inclusive, that innovation divorced from linguistic and cultural diversity leaves too many people behind. As millions of Africans begin interacting with AI in their mother tongues, they move from the periphery to the center of the global digital transformation.
