Abstract:

Amid growing global technological competition, understanding the reasoning behind U.S. artificial intelligence (AI) policy is crucial for evaluating its strategic impact and potential risks for China. Building on the theoretical framework of technological nationalism, this study introduces a four-dimensional analytical model covering national technological strategic positioning, the establishment of technological sovereignty, the connection between technology and national security, and the development of global technological discourse power.

Using the Dynamic Topic Model (DTM), we analyze 306 U.S. AI policy documents issued from 2015 to 2025 to identify key thematic trends and map the evolving policy emphasis. Complemented by keyword co-occurrence analysis, the study examines the interdependencies among policy themes to reveal how technological nationalism has become structurally embedded in the United States’ AI policymaking process and how this shapes its external technological stance toward China. The findings demonstrate that U.S. AI policy is characterized by a strong technological nationalist orientation. Domestically, the United States tightly links AI development with national security goals, creating a legitimacy narrative around “national security threats” to justify restrictive technological measures targeting China. Internationally, it promotes governance frameworks infused with American values and fosters alliances with like-minded partners, thereby institutionalizing a bloc-based regime of technological containment.

In response, the study proposes three strategic pathways for China—foundational breakthroughs, standard elevation, and partnership restructuring—aimed at mitigating external technological vulnerabilities and enhancing its strategic autonomy and competitiveness within the global AI landscape. These findings contribute to a deeper understanding of the geopolitical dimensions of AI governance and offer actionable insights for national technology strategy in an era of renewed techno-political rivalry.

Key words: U.S. artificial intelligence policy, techno-nationalism, dynamic topic modeling, text co-occurrence analysis

摘要：

全球科技竞争加剧背景下,深入剖析美国人工智能政策演化逻辑,可为中国化解外部技术风险提供重要启示。基于技术民族主义理论,搭建“国家技术战略定位、技术主权构建、技术与国家安全耦合、全球技术话语权塑造”四维分析框架。利用动态主题模型(dynamic topic model)识别2015-2025年306份美国人工智能政策文本的主题,动态追踪政策主题强度演化,并基于以上四维分析框架解读模型结果,再结合文本共现方法,对模型结果进行主题词共现分析,研究不同政策主题间的协同逻辑,揭示技术民族主义在美国人工智能政策中的体现及对华人工智能发展的影响。研究发现,美国人工智能政策具备鲜明的技术民族主义属性。为获得人工智能全球领导地位,美国对内将人工智能技术与国家安全议题深度绑定,以“国家安全威胁”为核心逻辑,构建对华技术管制的合法性叙事基础。对外输出嵌入美式价值观的技术规则,并组建认可美式价值观的技术联盟实现对华技术遏制阵营化。对此,提出“底层突围、标准跃升、伙伴重塑”三条路径,应对外部技术风险,提升中国人工智能领域的竞争主动权。

关键词: 美国人工智能政策, 技术民族主义, 动态主题模型, 文本共现分析