Publication
From Extraction to Gardening: Ubuntu-Inspired Provocations for Relational AI and CSCW
Ajit G. Pillai; Jaydon Farao; Houda Elmimouni; Cat Kutay; Friederike Fröbel; Wakanyi Hoffman; Hafeni Mthoko; Marly Muudeni Samuel; Alfred Malengo Kondoro; Eman Abulmagd; Nour Boulahcen; Katja Becker; Wanjiru Mburu; Martha Mosha; Chris Muashekele; Shaimaa Lazem
In: CSCW Companion '26: Proceedings of the 2026 Companion of the Computer-Supported Cooperative Work and Social Computing. Conference on Computer-Supported Cooperative Work and Social Computing (CSCW-2026), October 10-14, Salt Lake City, UT, USA, ISBN 979-8-4007-2378-0, Association for Computing Machinery, New York, NY, USA, 10/2026.
Abstract
This paper presents findings from a workshop examining how African indigenous philosophies can reshape Human-Computer Interaction and AI development. Drawing on discussions, storytelling activities, and thematic analysis of participant reflections gathered during a workshop in Egypt with researchers and practitioners from across Africa, we articulate a set of provocations for CSCW and AI research. We challenge dominant HCI paradigms rooted in Euro-American traditions that render African experiences invisible through "WEIRD" (Western, Educated, Industrialized, Rich, Democratic) frameworks. Drawing on Ubuntu philosophy’s emphasis on relational intelligence, "I am because we are" we propose shifting from extractive, individual-centric design toward community-oriented, dialogical approaches. Workshop participants identified fundamental disconnects between African realities and current AI architectures, advocating for "Small AI" models that are situated, sustainable, and locally governed rather than energy-intensive "God models." We present three provocations: redefining intelligence as relational stewardship relying on local knowledge rather than extraction to other contexts; acknowledging researcher positionality through "seamful design"; and achieving technological sovereignty through cooperative, right-sized AI ecosystems. Our findings suggest that centering African indigenous intelligence requires not merely dataset translation but fundamental counter-visualization of technology futures rooted in community exchange where data and its usage can be trusted.
