No Fate But What We Make: Structured Divergence in the Artificial Intelligence Governance Approaches of the U.S., EU, and PRC from 2016-2024

Abstract

The race to govern artificial intelligence has become a critical arena of great power competition. The United States, European Union, and People's Republic of China are each developing AI governance frameworks shaped by their distinct political systems, economic interests, and strategic priorities. Whether these frameworks will converge toward global standards or fragment into incompatible regulatory blocs carries significant implications for technological innovation, international trade, and geopolitical influence.

This dissertation employs computational text analysis to systematically compare AI governance approaches across these three jurisdictions and international organizations from 2016 to 2024. Using natural language processing methods, including topic modeling and semantic similarity analysis, I analyze 37 official policy documents to measure both lexical differences (word choice and terminology) and semantic convergence (shared concepts despite different wording).

The findings reveal an important pattern: "structured divergence." Different governance frameworks employ distinctive vocabularies reflecting their political cultures. The U.S. and EU emphasize risk management, China prioritizes innovation leadership, international organizations focus on societal impact, yet they increasingly share conceptual foundations, particularly regarding AI safety, ethical principles, and the need for standards. This disconnect between surface-level language and underlying meaning suggests that policy coordination may be more feasible than superficial differences imply.

The temporal analysis shows global AI governance evolving through three phases: innovation-focused optimism (2016-2017), growing attention to societal impacts (2018-2019), and recent emphasis on risk management and workforce effects (2020-2024). This evolution reveals governance responding reactively to technological developments rather than proactively shaping them.

These findings identify a middle path between global fragmentation and full harmonization: competing governance models can maintain political and cultural distinctiveness while achieving sufficient alignment on technical standards and procedural mechanisms to enable international cooperation. For policymakers, this suggests focusing on operational interoperability rather than demanding wholesale convergence. For researchers, this demonstrates the utility of computational methods for analyzing policy coordination in emerging technology domains beyond AI.

Disciplines

International Relations | Internet Law

Subject Area

Artificial intelligence; International relations; International law; Public policy

Department

International Relations (INR)

First Advisor

Bradizza, Luigi

Second Advisor

Raymond, Chad

Third Advisor

Altounian, David

Date of Award

2026

Document Type

Dissertation

Degree Name

Ph.D.

Link to Full Text

Share

COinS