Data as Public Infrastructure: Unlocking AI-Driven Government Services Through Trusted Data Ecosystems

Abstract

Governments increasingly seek to deploy artificial intelligence (AI) to improve public services, yet progress remains constrained by fragmented data governance, limited interoperability, and declining public trust. This paper advances the argument that reconceptualizing government data as public infrastructure governed through trusted data ecosystems provides a viable pathway for enabling AI-driven public services while safeguarding privacy, accountability, and equity. Drawing on a mixed-methods study of data-as-infrastructure implementations across 45 jurisdictions, the research integrates comparative policy analysis, technical architecture assessment, and interviews with 87 stakeholders. Findings identify five interdependent governance dimensions legal, technical, institutional, trust, and economic that distinguish effective data infrastructure from conventional data management approaches. Jurisdictions adopting infrastructure-oriented governance demonstrate significantly faster AI deployment, improved cross-agency data sharing, reduced duplicative data collection, and higher levels of citizen trust. The study contributes theoretically by linking infrastructure studies, institutional economics, and AI governance, and practically by providing evidence-based guidance for policymakers, government technology leaders, and civil society actors seeking to operationalize trustworthy AI in the public sector.