Participation Governance: Governing How Powerful Systems Participate in Human Society

This paper defines Participation Governance as a new governance framework for determining how, when, and under what relational conditions powerful systems should participate in human life. Existing approaches to AI governance primarily address capability, safety, legality, fairness, authorization, human oversight, and institutional accountability. Participation Governance complements these approaches by treating Participation itself as an independent object of design, control, evaluation, and learning. The paper distinguishes four governance questions: Capability — what a system can do;Safety — whether an action creates unacceptable harm;Permission — whether the system is authorized to act;Appropriateness — whether that capability should be exercised here, now, in this role, at this relational distance, intensity, and level of authority. It introduces five core components: Participation Modes Policy Arbitration Human Return Relationship Outcome Relationship Runtime Together, these components govern the transition through which system capability becomes socially situated and relationally consequential. The paper positions Participation Governance as a missing middle layer between system capacity and human consequence. It does not replace AI safety, alignment, human-centered AI, meaningful human control, or institutional governance. It provides a framework for coordinating them at the point where capability becomes participation. Although developed in response to increasingly capable and agentic AI systems, Participation Governance may also be applied analytically to other powerful actors, including families, educational institutions, organizations, medical systems, public institutions, and states. It proposes a common governance object, not a uniform governance regime. This paper serves as the category-defining work for Participation Governance within the broader Relationship-Aware AI Research program. Its normative foundation is presented in the companion paper: The Constitution of Relationship: A Normative Foundation for Participation Governance in Human–AI Society