摘要
[目的/意义]分析美国智库对深度求索(DeepSeek)崛起的看法,有利于把握美国对华人工智能政策的未来走向,为我国精准制定应对措施提供理论支撑。[方法/过程]本研究采用政策文本分析与观点聚类方法,系统剖析了兰德公司、卡内基国际和平基金会、美国信息技术与创新基金会和斯坦福以人为本人工智能研究中心4家美国权威智库有关DeepSeek的评论文章和研究报告。研究聚焦这些观点如何塑造美国未来人工智能政策,并评估其对中国人工智能发展的潜在影响。[结果/结论]研究发现,4家智库普遍认为美国应对DeepSeek的崛起保持警惕,并采取有效措施以维护其国家安全与技术优势。兰德公司等从国家安全视角出发,主张将芯片管制作为制约中国人工智能发展的战略手段;而以美国信息技术与创新基金会为代表的智库则基于商业利益考虑,主张美国应专注于保持并扩大其在人工智能领域的领先地位和市场占有率。基于此,我国应采取以下策略:一是,强化自主创新,突破技术瓶颈;二是,利用开源生态反制技术垄断;三是,建设自主可控的高质量国家级语料库;四是,推动构建公平的国际科技合作新秩序。
[Purpose/Significance]Analyzing U.S.think tanks’discourse on the rise of Chinese AI models like DeepSeek is helpful to clarify U.S.policy responses to China’s technological advances and provide theoretical support for China to formulate targeted countermeasures.[Method/Process]This study applies policy text analysis and thematic clustering methods to analyze commentaries and research reports from four influential U.S.think tanks—RAND Corporation,Carnegie Endowment for International Peace(CEIP),Information Technology&Innovation Foundation(ITIF),and Stanford Institute for Human-Centered Artificial Intelligence(HAI).It investigates how these perspectives shape prospective U.S.AI policies and assesses their potential impact on China’s AI advancement.[Result/Conclusion]The findings reveal a shared concern among the four think tanks over Deep-Seek’s rapid ascent,indicating that the U.S.is likely to adopt strategic measures to safaguard national security and maintain technological leadership.RAND and similar institutions,focusing on national security,advocate chip-restriction policies controls as a means to constrain China’s AI progress.By contrast,ITIF emphasizes commercial interests,urging the U.S.to concerntrate on sustaining and expanding its market share and technological dominance in AI.Drawing on the policy analysis,we recommend the following strategies:(1)Strengthen independent innovation to address critical technological bottlenecks;(2)Leverage open-source ecosystems to reduce systemic dependencies on monopolized technologies;(3)Develop a nationally curated,high-quality AI training corpus;(4)Lead international dialogues on balanced tech cooperation frameworks.
作者
蒋甜
钱力
常志军
Jiang Tian;Qian Li;Chang Zhijun(National Science Library,Chinese Academy of Sciences,Beijing 100190;Key Laboratory of New Publishing and Knowledge Services for Scholarly Journals,National Press and Publication Administration,Beijing 100190;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190)
出处
《智库理论与实践》
北大核心
2026年第1期144-152,共9页
Think Tank:Theory & Practice
基金
国家社会科学基金项目“AI4S科技文献知识底座的理论体系及建设方法研究”(项目编号:24BTQ043)的研究成果之一。
关键词
DeepSeek
人工智能
美国智库
国家安全
商业利益
芯片管制
DeepSeek
artificial intelligence(AI)
U.S.think tanks
national security
commercial interests
chip restrictions
