Multilateral Coordination for the Proactive Governance of Artificial Intelligence Systems

Abstract

The existing international governance of artificial intelligence (AI) is fragmented. With over 190 national jurisdictions and numerous regional or thematic bodies, we lack a universal entity charged with coordinating a unified response to the risks emanating from this technology. This working paper proposes the creation of a multilateral entity with the express mission of mitigating the shared large-scale high-risk harms caused directly or indirectly by AI systems. Its agenda is to identify vectors of risk, coordinate global action, and seek the enforcement of agreed-upon rules. Its design includes elements needed for the global governance of AI such as: concrete mechanisms to incentivize member state participation, an organizational structure that recognizes the need for engagement with multiple and diverse levels of authority, and a timeline for its creation.

Keywords: International governance of artificial intelligence, multilateral coordination, general purpose AI systems, narrow AI, enforcement of governance