Informatics Seminar Series
Spring Quarter 2021

Abstract:
We consider designing AI in indigenous contexts in a collaboration between Ju|’hoansi people and researchers associated with the International University of Management , Namibia and Cambridge . Probabilistic programming languages offer potential for new predictive tools to suit settings where big data is not available because of ethics, practicalities or its rarity. Scientists have included indigenous people’s knowledge in models that apply Bayesian inference to predict animal and plant patterns. Our aim, however, is for local inhabitants to use such tools driven by their own aspirations and knowledge practices. So far our work has situated discussions of mathematics in everyday reasoning about social, ecological, and other phenomena in the Kalahari. Our conversations, observations, games and stories about probability draw attention to predictive logics that emerge in knowledge that is, as Tim Ingold (2007, 2011) would say, “alongly integrated” and formed as people continuously wayfare in the world they inhabit. Predicting in practices that integrate and embody knowledge along paths differs from predicting by upwardly integrating discretely sampled data and calculating frequency amongst aggregates. My talk will discuss our current analysis of the translations that occur when we seek to encode predictive practices in mathematical statements.

Bio:
Nic Bidwell has researched at the intersection of Human-Computer Interaction (HCI) and social informatics with a focus on the Global Souths for 17 years. This encompasses working with Aboriginal people in far north Australia and inhabitants of rural Namibia, South Africa, Mozambique, Kenya, Uganda, Indonesia, Argentina, Mexico and India. She initiated the first panel on Indigenous Led Digital Enterprise at a leading HCI forum in 2008, co-founded the African HCI Conference in 2016 and her ethnographically informed design work with rural collaborators in South Africa received an award for its contribution to social and economic development. Nic’s analyses of relations between spectrum regulation and involvement in designing and deploying technologies for community networks (CNs) and community radio have also informed policy debate. Her recent studies include experiences of conditional programming in philanthropic donations, interactions with algorithms in ride-sharing platforms, and onto-epistemic translations in designing probabilistic programming languages for marginalized knowledge practices. Nic has taught in universities around the world and, for the past twelve years, lived in southern Africa where she is an adjunct Professor at the International University of Management, Namibia.