Organisational Knowledge Brokerage in Science Diplomacy: The Making of AI Governance at UNESCO and the European Union

Abstract

We examine how expertise, evidence and authority are organised in the international governance of Artificial Intelligence (AI), and what this reveals about the practice of science diplomacy. Science-diplomacy scholarship has become increasingly attentive to the political character of expertise and to the ways knowledge is mobilised within international governance. Yet less attention has been paid to the organisational practices through which such knowledge is brokered within governance processes. Drawing on documentary analysis and 23 semi-structured interviews, we compare UNESCO’s Recommendation on the Ethics of Artificial Intelligence and the EU AI Act. Through an abductive analysis, we develop Organisational Knowledge Brokerage (OKB) to explain how a contested object such as AI is made workable for governance. We identify four connected practices through which this occurs: stabilising the governance object, legitimising expertise, building authority and operationalising governance. These practices took different forms under contrasting institutional mandates and governing capacities. UNESCO worked through global intergovernmental standard-setting and implementation support, while the EU combined supranational regulation with expert advice and Member State coordination. Across both cases, OKB also produced a cross-cutting process that we term normative translation, through which contested values acquired practical meaning in definitions, assessments and procedures. We show how this brokerage allowed each organisation to establish workable and revisable settlements around AI under conditions of technological uncertainty and dispersed authority. The article therefore develops a practice-oriented account of science diplomacy as organisational work through which knowledge, values and institutional authority are assembled for collective action in international technology governance.