Overview

Many interventions have been developed, such as tools, guidelines, frameworks, and policies, that aim to improve responsible AI (RAI) practices across domains. The design and evaluation of effective RAI interventions require thoughtful collaboration among practitioners and researchers from diverse backgrounds, disciplines, sectors, and lived experiences. Although there is a growing body of recent CSCW work on RAI, as well as research exploring RAI issues and approaches in social computing contexts, there has not been a designated space for CSCW researchers to explore and learn from each other about collaboratively designing and evaluating RAI interventions.

To this end, we are conducting a two-part SIG session at CSCW 2024, consisting of a series of lightning talks and workshop activities. Our SIG aims to facilitate CSCW researchers learning from each other's shared challenges and successes around the design and evaluation of RAI interventions in diverse domains and contexts, creating synergy and amplifying the overall impact of CSCW research in the area of RAI. Topics we would like to discuss inlude but not are limited to:

What might it mean to evaluate RAI interventions? What does a successful RAI intervention look like, and who determines success?

How should we design and evaluate RAI interventions in domain-specific social computing systems where AI is used?

How might the research methods within the CSCW research community be leveraged to support more effective collaboration in designing and evaluating RAI interventions? What, if anything, might need to change about those research methods when applied to RAI?

How might CSCW research help bridge the gap between governance mechanisms and practice (e.g., EU AI Act, AI Risk Management Framework)?

How might CSCW researchers better communicate and collaborate with researchers and practitioners in other domains, such as policymakers, industry AI teams, and government agencies, to explore topics around RAI interventions?