Concept note: Embodied Data and Just AI – An Exploration of Key Questions

Anja Kovacs
05 Apr 2024

Current work on data justice highlights that, while data-driven technologies could be harnessed to support Africa’s development, at present, they often serve to further undermine the continent’s geopolitical position and harm its people. Data-related practices reinforce historical imperialist processes, including resource extraction, the imposition of Western values, the misrepresentation of African people and particularly those at the intersections of multiple inequalities, and the misappropriation of indigenous knowledge systems. The concept of data justice has allowed us to speak about these complex harms in order to arrive at data governance policies that further just outcomes.

This research seeks to further develop this body of work by taking as its starting point that data about people can not simply be approached as a resource. Instead, it argues, our thinking needs to take into account that such data, and the effects of its manipulation, are, in fact, embodied. Adopting a feminist approach to data, particularly one that understands data in contemporary life as integral to our embodied or “lived” experiences, sheds new light on how data and its uses perpetuate injustice. This research seeks to map these injustices specifically in the context of AI. This, in turn, empowers us to forge further alternative visions of AI that challenge, subvert, and redress the harms that AI currently often creates and amplifies.