In 2024, I joined Research ICT Africa (RIA) as a Senior Fellow, and worked with a team of RIA researchers to explore how thinking about data as embodied influences and advances thinking about Just AI. Adopting a feminist approach to data, we took as our starting point the understanding that data in contemporary life is integral to Africans’ embodied or “lived” experiences of their bodies. Taking such an approach, we hypothesised, would then shed new light on how data and its uses perpetuate injustice in the context of artificial intelligence (AI), and help us to determine what needs to be done to turn this around in Africa as elsewhere.

This collection brings together some of the key reflections this exercise sparked in each researcher’s mind.

In the introductory essay, I sketch out an overarching framework to guide thinking about data as embodied in the context of AI. I first discuss how dominant narratives that conceptualise data as a resource erase people and their agency, and why this is problematic. I then explain why it is essential to bring bodies and embodiment back into data and AI governance. I expand on the paradigm shift our bodies are undergoing as a result of datafication, and outline how the recent entanglement of our bodies with data produces not only a different kind of body, but also different forms of embodiment, or different ways of being human. Following Irma van der Ploeg (2012), I argue that both these shifts are not of a representational nature, but an ontological one. Finally, I describe how an understanding of data related to people as embodied will shift our thinking about AI. I do so by re-examining two key moments in the AI life cycle through the lens of the shifts outlined earlier: the collection of embodied data to create and feed AI (inputs), and the embodied effects of AI applications (outputs). As will become evident, for AI to be just, it is essential that the learnings of this re-examination be taken into account.