DIKIW逻辑链下GPT大模型对文献情报工作的潜在影响分析  

Analysis of the Potential Impact of GPT Large Model Under DIKIW Logic Chain on Documentation and Information Services

作　　者：刘细文[1,2] 孙蒙鸽 王茜 付芸[1,2] 王燕鹏 郑新曼 Liu Xiwen;Sun Mengge;Wang Xi;Fu Yun;Wang Yanpeng;Zheng Xinman(National Science Library,Chinese Academy of Sciences,Beijing 100190;Department of Information Resources Management,School of Economics and Management,University of Chinese Academy of Sciences,Beijing 100190)

机构地区：[1]中国科学院文献情报中心,北京100190 [2]中国科学院大学经济与管理学院信息资源管理系,北京100190

出　　处：《图书情报工作》2023年第21期3-12,共10页Library and Information Service

基　　金：国家自然科学基金重点项目“数智赋能的科技信息资源与知识管理理论变革”(项目编号:72234005)研究成果之一。

摘　　要：[目的/意义]以ChatGPT为代表的GPT类生成式大模型发展日益蓬勃,引起学界和业界的广泛讨论,同时也对文献情报工作产生了潜在的难以估量的影响。以DIKIW逻辑链为指导框架,研究GPT大模型对文献情报工作的潜在影响。[方法/过程]以DIKIW为主线,梳理全数字化文献情报工作概念框架;在此框架下分析、评估GPT类大模型对文献情报工作的影响,并构想GPT影响下的文献情报工作流程。[结果/结论]研究认为,GPT类大模型对文献情报工作的本质影响是打通DIKIW链条中的“数据→信息→知识”部分,即原本“将原始数据加工为信息、进而增值为知识”的人工操作步骤,将由GPT类大模型在内的AI工具链条式完成。因此,文献情报工作需要从“大规模加工机器可读可理解的数据”“大规模集成情报开源模型”等方面进行突破改进,才能更好地支撑场景驱动的智慧化情报服务工作。[Purpose/Significance]The development of GPT generative large language models represented by Chat GPT has become increasingly vigorous,which has aroused extensive discussion in the academic circles and the industry,and has had an immeasurable impact on documentation and information service.[Method/Process]Firstly,by taking DIKIW logic chain as the main line,this article sorted out a set of conceptual frameworks of digitized documentation and information service.Then,under the framework,it analyzed and evaluated the impact of GPT models on it,and conceived the corresponding workflow under GPT.[Result/Conclusion]It concludes that the essential impact of GPT large models on documentation and information service is to break through the“data→information→knowledge”part of the DIKIW,that is,the manual operation of“processing raw data into information and knowledge”will be completed by AI tools including GPT large models.Therefore,the documentation and information service will make progress and improvement in“large-scale processing of machine-readable and understandable data”and“large-scale integrated intelligence open source model”,so as to better support scenario-driven intelligent intelligence services.

关 键 词：数字化文献情报工作 GPT 大语言模型 知识架构师 情报工作生态 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] G25[自动化与计算机技术—控制科学与工程]