AI Chatbots Can Boost Public Health in Africa—But Only If They Speak Our Languages
AI-powered health chatbots are rapidly emerging across Africa, promising to revolutionize access to health information—but there’s a catch: most don’t speak local languages.
A new study by researchers from the University of Cambridge has found that while artificial intelligence is advancing in Africa’s public health space, its impact remains limited by one critical factor—language inclusion.
The research team, which reviewed nearly a decade of academic studies, discovered that although AI tools like chatbots and sentiment analysis systems are being developed to improve healthcare, only 4% of these tools have demonstrated real-world health outcomes. The rest remain confined to the lab.
Most of these tools operate in English and French—languages associated with colonial history—leaving large swathes of the African population unable to access or act on the health information they provide.
“The health benefits of AI won’t reach ordinary Africans if the tools only speak in colonial languages,” said Dr. Anna Barford, a senior researcher involved in the study.
One standout exception was a chatbot built by researchers from the University of Chicago and the Busara Center for Behavioral Economics. Designed for Facebook Messenger and targeted at vaccine-hesitant users in Kenya and Nigeria, the English-only chatbot increased people’s willingness to get vaccinated by up to 5%—a measurable public health impact.
The project’s success was largely due to deep local engagement. The developers conducted focus groups and interviews in both countries to understand cultural concerns before building the tool.
A Gap in Local Relevance
The authors argue that most AI language tools in Africa remain experimental, with developers focusing more on the tech than on community needs. And while large language models like GPT-4 have made development easier and more accessible, the real challenge now is making them locally relevant.
“There’s a real opportunity for governments, startups, and donors to invest in AI systems built with African languages in mind,” said Anna Korhonen, a professor of Natural Language Processing.
Involving communities, training models in indigenous languages, and embedding these tools in health systems could accelerate their public health impact.
The researchers emphasize that future AI for health efforts must move from lab to life—with rigorous field testing, real-world deployment, and measurable outcomes. Without that, Africa risks falling into a pattern where powerful technologies are built for the continent, but not with it.
In other news:Kenya Cannot Sustain Free Education Anymore, Treasury CS Mbadi Declares
AI Chatbots Can Boost Public Health in Africa—But Only If They Speak Our Languages