人工智能赋能智库建设——美国国际战略研究中心的启示  

AI Empowering Think Tank Development—Insights from the Center for Strategic and International Studies(CSIS)

作　　者：蒲云强 唐川 敖丽娟 Pu Yunqiang;Tang Chuan;Ao Lijuan(National Science Library(Chengdu),Chinese Academy of Sciences,Chengdu 610299;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences Beijing 100049)

机构地区：[1]中国科学院成都文献情报中心,成都610299 [2]中国科学院大学经济与管理学院信息资源管理系,北京100049

出　　处：《智库理论与实践》2025年第2期93-105,共13页Think Tank:Theory & Practice

基　　金：中国科学院成都文献情报中心创新基金青年项目“基于大模型和多源数据的科技安全前瞻情报研究”(项目编号:E3Z0000802);中国科学院战略研究专项项目“‘双碳’动态监测和信息领域科技制高点研究”(项目编号:GHJ-ZLZX-2024-07)研究成果之一。

摘　　要：[目的/意义]人工智能(Artificial Intelligence,AI)技术的快速发展为智库研究带来了新机遇。AI技术赋能使智库能够更高效地处理复杂的研究任务,推动其从传统研究模式向智能化、动态化的全面转型。借鉴国外智库在AI技术布局、方法创新与实际应用方面的实践经验,有助于提升我国智库的智能化水平和国际竞争力。[方法/过程]鉴于美国国际战略研究中心(Center for Strategic and International Studies,CSIS)在AI领域的前沿探索及其公开的研究过程,通过文献调查与案例分析相结合的方法,探讨CSIS在项目管理、外部合作和人员配置中的AI布局,并分析其典型研究报告,以梳理大语言模型在其研究分析中的具体应用。[结果/结论]CSIS率先探索AI在智库建设中的应用,通过将研究与实践相结合,系统布局并开展专题试点,涵盖研究报告分析、AI生成图像和音频等多个领域,提升了研究方法的多样性和成果的影响力。我国智库应围绕AI技术优化人才配置,深化对外合作模式,并丰富分析手段,加快开源化进程,积极探索大语言模型和多模态等AI技术的应用,抢占AI赋能的先机,推动智库智能化发展。[Purpose/Significance]The rapid development of Artificial Intelligence(AI)technology has opened new opportunities for think tank research.By leveraging AI,think tanks can more efficiently address complex research tasks,driving a comprehensive transformation from traditional research models to intelligent and dynamic approaches.Drawing from foreign think tanks'practical experiences in AI technology deployment,methodological innovation,and real-world applications helps enhance the intelligence level and international competitiveness of China's think tanks.[Method/Process]Focusing on the Center for Strategic and International Studies(CSIS)as a leading example in AI exploration,this study adopts a multi-source information integration and case analysis approach.It examines CSIS’s AI deployment in areas such as project management,external collaboration,and personnel allocation,while analyzing representative research reports to identify the specific applications of large language models in its research and analytical processes.[Result/Conclusion]CSIS has pioneered the integration of AI into think tank development by combining research with practice,systematically organizing and implementing thematic pilots projects.These initiatives encompass diverse applications,including research report analysis,AI-generated images and audio,thereby enhancing methodologies diversity and amplifying the influence of research outcomes.To advance think tank intelligence,Chinese think tanks should prioritize optimizing talent allocation for AI,deepening models of external collaboration,diversifying analytical methods,accelerating open-sourcing initiatives,and actively exploring the applications of AI technologies such as large language models and multimodal approaches.By seizing the opportunities presented by AI empowerment,Chinese think tanks can position themselves at the forefront of intelligent development.

关 键 词：美国国际战略研究中心 人工智能 大语言模型 多模态应用 智库建设 

分 类 号：G359[文化科学—情报学]