Infrastructure, Industrial Policy, and the Political Economy of AI

Session Submission Type: Panel

Abstract

This panel brings together four papers that collectively investigate the governance of artificial intelligence (AI) through a conceptual differentiation in core elements of its underlying infrastructure: physical, technological, methodological, and computational. Together, the panel engages with different understandings of the systems which enable science and technology policy by investigating how upstream institutional contexts, resource decision-making processes, and computational tactics shape AI governance and innovations.

This objective is managed by organizing the panel within two complementary approaches to AI research: procedural and content. The first and second papers focus on the processes behind AI governance and innovation by focusing on physical and technological infrastructure respectively. The first paper develops a comparative institutional framework to explain divergent national approaches to data center governance in both South Korea and Australia. The study combines topic modeling, network analysis, and qualitative discourse analysis to examine regulatory documents on data center infrastructure and asset management, finding a structural governance gap in determining how new technologies are categorized, prioritized, and managed within the policy process. The second paper shifts to examine the technological core of AI, questioning how the governance control/openness of capability for computational hardware design shape firm-level innovation incentives. While the first paper demonstrates how pre-existing logics enable path dependent responses to AI governance, the second paper provides insight into the ecosystem shocks which can change the trajectory innovation of the critical technology underlying AI development.

Given this foundation, the third and final papers round out an understanding of the relationship between governance systems and technology science by focusing on innovative approaches to AI use. The third paper leverages in-depth interviews with academic researchers working in manufacturing and materials sciences across U.S. states to draw conclusions on how AI transforms not only what science produces but also the underlying standards of methodological inquiry, such as the infrastructure behind standards of validation, reproducibility, and collaboration in research design. Finally, the fourth paper applies this understanding and introduces a novel computational methodology using AI – “artificial intuition” – that enables scalable, reproducible infrastructure for comparative portfolio analysis among countries, such as the U.S. and China in this study, who actively compete for global scientific and technological leadership.

The panel engages with APPAM’s 2026 conference theme by mapping the logic of policy variation across national and subnational governments onto a global perspective of AI innovations. The panel connects a core insight into policymaking in a federal context – how governance outcomes are shaped by institutional variation – to AI policy and management by exploring evidence from South Korea, Australia, China, Europe, and the U.S. states. In this way, the panel contributes to theoretical insight on AI governance by focusing on the conceptual differences in AI systems, including infrastructural and institutional alignment. Methodologically, it integrates studies leveraging quantitative and qualitative approaches to enable a holistic examination of key challenges in implementing and managing innovation. Practically, it introduces globally comparative and cross-sectoral perspectives on science policy, enriching a discussion about how various actors in society and different nations approach AI governance and capabilities.

Policy Area

• Innovations in Science and Technology

Chair

• Emily Boykin, Augusta University

Discussant

• Zeewan Lee, National University of Singapore – Lee Kuan Yew School of Public Policy

Organizer

• Emily Boykin, Augusta University

Individual Presentations

• 1. How Institutional Defaults Shape Data Center Regulation in Australia and South Korea - Non-Presenting Co-Author: Seokkyun Woo, Korea Advanced Institute of Science and Technology; Presenting Author: Jeonghyun Seo, Korea Advanced Institute of Science and Technology

• Innovation in AI accelerator platforms: patent-based analysis of innovation dynamics to inform goverance - Presenting Author: Anurag Panda, ETH Zurich

• Artificial intelligence and scientific infrastructures: Effects on resources, skills, and communications - Presenting Author: John Paul Nelson, Oregon State University

• Mapping the Science Race: Comparison of the US and Chinese Technology Portfolios using Artificial Intuition - Presenting Author: Alexandra Graddy-Reed, University of Southern California