Collaborative Approaches to AI Governance: Exploring Co-Design and Co-Regulation Models

Abstract

As Artificial Intelligence (AI) systems become increasingly complex and their social impacts more profound, traditional methods to technology governance are proving inadequate. Neither top-down regulation nor unfettered market freedom appears capable of addressing the multifaceted challenges posed by AI technologies. In response, participatory and democratic strategies are gaining traction in AI governance discussions as potential solutions. However, the practical implementation of these approaches and the challenges they may face remain underexplored in the literature, with existing theories often remaining idealistic and detached from real-world constraints. This dissertation introduces a novel approach to AI governance by integrating co-design and co-regulation methodologies. It addresses three key research questions: (1) How might co-design and co-regulation be defined and integrated in the context of AI governance? (2) How can domain-specific expert knowledge be effectively elicited and integrated into AI governance policies? (3) How do co-regulation models in related domains facilitate stakeholder collaboration, and what lessons can be applied to AI governance? The research methodology combines theoretical analysis with empirical research, featuring two in-depth case studies. The first study examines co-design in AI systems providing legal advice, engaging legal experts to develop guiding principles based on time-tested wisdom in legal professional communities. The second investigates co-regulation in online content moderation in South Korea, comparing web comics and news industries to identify critical factors for successful co-regulation. Both studies apply established frameworks to contexts previously unexplored in the AI governance literature. The key findings reveal that effective participatory governance requires strategic calibration of participation rather than maximizing participation, balancing diverse needs, motives, and contextual factors. This nuanced approach acknowledges that while multi-stakeholder involvement is crucial, shared responsibilities can lead to diffused accountability. This work specifies key questions for utilizing participatory or collaborative methods in AI governance and identifies significant contextual factors that may determine the success of such systems. It equips policymakers, AI developers, academics, and affected communities with pragmatic viewpoints, moving beyond abstract ideals to address real-world complexities. As a starting point for future research at the intersection of participatory design, collaborative regulation, and AI development, this dissertation challenges the field to advance from theoretical discussions to pragmatic, context-sensitive strategies in AI governance.