Efficiency at What Cost? The Political Dynamics of AI-Driven Reforms in the U.S. and China

• In Event: Governance in Action: Comparative Studies of Bureaucracy and Policy

Abstract

This study compares how AI-driven governmental reforms in the United States and China reflect and reshape political power amid complex governance structures. In the U.S., Elon Musk’s Department of Government Efficiency (DOGE) claims to streamline operations through staff reductions and agency consolidations, yet critics question whether these changes are motivated by genuine efficiency concerns or partisan objectives. Legal challenges and political controversies underscore the tension between innovation and institutional accountability.

Meanwhile, China’s state-directed approach integrates AI tools, such as DeepSeek, into public administration, allowing for top-level orchestration of technological adoption. Initial findings indicate improved service delivery, but broader implications for transparency and civil liberties remain contested. By framing these reforms as efficiency-driven, both governments seek to legitimize their initiatives, even as they confront domestic debates over the concentration of authority.

Employing a comparative case study method and insights from political theories on power and institutional change, this research investigates how divergent systems deploy AI as an instrument of governance. Policy documents, media discourse, and expert interviews reveal that AI adoption can either reinforce existing hierarchies or challenge them, depending on leadership intent and legal-institutional constraints. The findings highlight how top-down directives and legal frameworks mediate these transformations, clarifying the political trade-offs that accompany the pursuit of efficiency.

Ultimately, this analysis underscores that power dynamics lie at the heart of technological reforms, raising fundamental questions about the conditions under which AI can genuinely serve the public good in diverse political contexts.

Author

• Chao Song, University of Chinese Academy of Social Sciences