Fostering African Data Commons

Abstract This research envisions an African data commons regime, rooted in the philosophy of Ubuntu, where training data for AI is pooled and shared as a common-pool resource. Artificial Intelligence (AI) is gaining traction in Africa but it is still in its nascent phase. While investment in Africa’s AI infrastructure has become a battleground in the U.S.–China race for influence, Africa suffers from a scarcity of high-quality training data. Currently, data governance centers on the notion of “privacy,” emphasizing the rights of individuals with respect to their personal information, resulting in data that is predominantly controlled by companies located outside of Africa, a phenomenon some describe as “data colonialism.” To foster a localized, African AI data ecosystem, the article proposes constructing a data governance regime rooted in the essence of collective responsibility. This approach fosters data sharing, given the intertwining of the philosophy of Ubuntu with the governance of commons. To highlight the importance of the infrastructural component of data governance, the article, grounded in Nobel Laureate Elinor Ostrom’s theories of the commons, proposes to treat training data as a common-pool resource. In this view, the analysis of various data commons will center on the physical components of data governance, specifically through the lens of the Institutional Analysis and Development (IAD) framework. The goal is to provide an analytical framework to decide what must be undertaken to cultivate a localized data ecosystem, drawing upon a governance system deeply embedded in an African worldview.