Participatory Design of AI Systems: Opportunities and Challenges Across Diverse Users, Relationships, and Application Domains

Abstract

Participatory design (PD) for Artificially Intelligent (AI) systems has gained in popularity in recent years across multiple application domains, both within the private and public sectors. PD methods broadly enable stakeholders of diverse backgrounds to inform new use cases for AI and the design of AI-based technologies that directly impact people's lives. Such participation can be vital for mitigating adverse implications of AI on society that are becoming increasingly apparent and pursuing more positive impact, especially to vulnerable populations. This panel brings together researchers who have, or are, conducting participatory design of AI systems across diverse subject areas. The goal of the panel is to elucidate similarities and differences, as well as successes and challenges, in how PD methods can be applied to Artificially Intelligent systems in practical and meaningful ways. The panel serves as an opportunity for the HCI research community to collectively reflect on opportunities for PD of AI to facilitate collaboration amongst stakeholders, as well as persistent challenges to participatory AI design.