Africell and Oxford University have partnered to study the economic and social impact of AfriGPT, an innovative SMS-based AI service, in West Africa. The research project, funded in part by Schmidt Sciences’s $3 million ‘AI at Work Program’, will investigate how AfriGPT is used in Sierra Leone and The Gambia, with a focus on its potential applications in job search, entrepreneurship, and education.
AfriGPT allows users to interact with ChatGPT via SMS, requiring only 2G mobile connectivity, making it accessible to those with limited internet access. This service opens up the potential advantages of AI chatbots to a poorer, younger, and more rural demographic.
The research aims to understand the behaviours, social and economic conditions, and other factors that affect AI use in Africa, to inform the development and distribution of AI tools that meet the needs of low-income and rural communities.
“We want to understand how AI chatbots are used in Sub-Saharan Africa, how this is different from global use, and what the benefits and risks of AI in this context are,” say Johanna Barop and Joseph Levine, DPhil researchers at Oxford University.
The findings of this research are expected to have significant implications for AI development in Africa, potentially informing strategies to promote responsible AI adoption in the region. By understanding how AI tools are used in low-income and rural contexts, developers can design more inclusive and effective AI solutions that address the unique challenges faced by these communities.
“Tools such as ChatGPT are increasingly taken for granted in Europe and North America. However, for both economic and technological reasons, they are much less common in Africa,” says Sam Williams, Africell’s Group Communications Director. “AfriGPT is promising because it enables mobile users in countries such as Sierra Leone, where internet penetration remains relatively low, to use AI chatbots without the need for internet access.”
The research is expected to yield valuable insights into how AfriGPT is employed in different regions, on different types of devices, and at different times of year, ultimately contributing to the development of AI tools that are tailored to the needs of African users.
