Online ReporterCassava Technologies, the pan-African technology group founded by tech entrepreneur, businessman and philanthropist Strive Masiyiwa, has joined a global coalition seeking to make artificial intelligence accessible to billions of people whose languages remain poorly supported by AI systems.
Cassava is among signatories spanning an initial 60 technology companies, AI laboratories, researchers, governments, philanthropies and civil society organisations backing a five-year goal to enable an estimated 3.4 billion people who speak languages underrepresented in current AI models to use AI tools in their own language and voice.
The organizations include Nvidia, Google, Microsoft, OpenAI Foundation, World Bank Group, UK Foreign, Commonwealth and Development Office, Amazon, Agentic AI Foundationand the Centre for Open Societal Systems (COSS) among many others.
The initiative, announced by the Gates Foundation this week, comes as artificial intelligence rapidly expands into education, healthcare, agriculture, financial services and public services, while much of the world’s linguistic diversity remains poorly represented in the technology underpinning the AI boom.
Of the world’s roughly 7 000 languages, only a small proportion have the digital text, speech data, tools and benchmarks needed to support sophisticated AI systems. Thousands are considered “low-resource languages”, leaving their speakers more likely to encounter AI that struggles to understand their words, accents, dialects and cultural context.
The resulting divide is not simply about internet access or computing power, but whether people can interact with increasingly important digital technologies in the languages they use every day.
Across the Global South, farmers, teachers, students and health workers may already have access to information that could improve livelihoods and services. But language barriersoften prevents the flow of that information between communities and digital systems, limiting the potential benefits of AI.
The coalition says the challenge extends beyond translating written text to building AI systems that can understand, interpret and respond accurately to the languages, dialects and cultural contexts in which people communicate.
Voice interfaces could be particularly important in communities where typing is less practical or where text-based services remain difficult to access. At the same time, dialects, slang, idioms and cultural context can change meaning, creating risks of misunderstanding (or hallucinations) when AI is used for information relating to health, education, agriculture or public services.
“AI needs to work well enough for speakers to accomplish what matters to them, whatever language they speak,” the Gates Foundation and its partners said in their joint commitment.
The coalition is calling for an open language layer built around responsibly collected and governed language data that AI developers can use under open licences. It also plans to develop benchmarks to measure progress and support the development of AI applications in underrepresented languages, while protecting privacy, consent and data sovereignty.
“We need to gather quality data responsibly and with consent, make non-proprietary data open to every builder, set benchmarks, and ensure model-makers use them,” the organisations said.
They called for investment in local research and technology ecosystems so communities contributing their languages are not merely sources of data, but participants and beneficiaries of the AI economy.
The challenge is particularly evident in Africa, home to thousands of languages, many of which have limited digital resources. Improving AI capabilities in these languages could become an important part of the continent’s wider digital transformation and efforts to broaden access to digital services.
But language technology alone will not close the digital divide. Reliable connectivity, electricity, affordable devices and digital skills will also determine whether advances in AI and voice technologies reach the communities they are intended to serve.
“No single company, country, or foundation can do this alone,” the organisations said. “It takes the AI labs that build models, the researchers who understand the languages, the organizations that serve communities, and the governments and funders willing to invest in (the) public good.”
For Cassava Technologies – which has taken the lead in harnessing and leveraging the power of AI compute across Africa through its subsidiary, Cassava AI – participation places the African technology group among organisations seeking to shape a more inclusive next phase of the AI revolution.
The coalition’s five-year ambition is ultimately more thansimply adding languages to AI systems but about enabling billions of people to learn, work, access essential services and participate in the digital economy without having to leave their own language behind.
