A Multi-level Collaborative Governance Framework for Designing Accountable AI Systems

Yu-Che Chen, University of Nebraska at Omaha, will present "A Multi-level Collaborative Governance Framework for Designing Accountable AI Systems" as part of the CPR Seminar Series.

Abstract:

How do we design accountable artificial intelligence (AI) systems? Our NSF-funded project team takes an interdisciplinary approach to develop and implement a multi-level governance design framework to answer the question. The public policy and service context is collaborative emergency management between tribal nations and U.S. governments. We implement a co-design of a custom-built AI chatbot for emergency management in indigenous communities and study the impact of such co-design on accountability. We strive for a culturally responsible way to engage with indigenous people for the AI chatbot co-design and research. We follow an interpretive approach in collecting and analyzing data. The project's preliminary findings suggest the importance of culturally responsible engagement, embedded multi-level governance, tribal sovereignty, participatory design, and explainable AI in designing a more accountable AI system.