Information Studies Colloquium: Participatory AI with Future Doctor Kelly Wagman

Please join us by Zoom for this week’s Information Studies Colloquium, Thursday, January 16, 2-4 p.m. The presenter will be Kelly B. Wagman, discussing participatory AI.

Bio:

Kelly B. Wagman is a PhD candidate at the University of Chicago in the Computer Science department where her focus is human-AI interaction. Kelly’s research centers around understanding how AI integrates with sociotechnical systems and how to improve the design of those systems. Previously she worked at Microsoft Research and received a MS in Comparative Media Studies from MIT and BAs in Computer Science and Economics from Brown University.

Abstract:

“Participatory AI” has been proposed as a mechanism for designing ethical and inclusive AI systems. However, scholars differ on theoretical definitions for participatory AI and few practical examples exist of participatory AI implementations in real-world settings. In this project, we develop a participatory AI system for climate that is being deployed in a Chicago neighborhood in order to understand the benefits and limits of this approach. More specifically, we partnered with climate scientists and a Chicago community non-profit organization to understand how AI models of microclimates can benefit community residents and incorporate their perspectives. To do this we conducted interviews, co-design workshops, and ethnographic work. Based on this qualitative work, we built a climate reporting and visualization app for the community to contribute data (e.g. photos of basement flooding) to the science AI pipeline as well as understand how these climate models can be made actionable in their own lives. Ultimately this project demonstrates that it is possible to develop a practical participatory AI system that benefits multiple stakeholders, although it is time-intensive and questions remain as to the scalability of such an approach.