人机合作:探索生成式AI在质性研究中的应用图景  

“Human-Machine”Collaboration:Exploring the application landscape of generative AIinqualitative research

作　　者：朱逸 阙明坤[3] ZHU Yi;QUE Ming-kun(Management College,Shanghai Sanda University;International Governance Research Center for Cyberspace,Fudan University;College of Education,Zhejiang University)

机构地区：[1]上海杉达学院 [2]复旦大学网络空间国际治理研究基地 [3]浙江大学国家高端智库教育学院

出　　处：《科学与社会》2025年第1期115-132,共18页Science and Society

基　　金：国家社科基金一般项目“增权理论视角下互联网‘零工群体’劳动权益保障研究”(23BSH057)。

摘　　要：生成式AI介入质性研究日益增多。学界就人工与机器在质性研究中的实践状况进行了诸多比对研究,发现生成式AI对于质性研究有积极的作用。本文选取中文环境下的家庭教育访谈文本作为文本基础,运用本土生成式AI,比对在质性研究(扎根理论、主题分析)中人工与机器的分析质量。研究采取人工与机器对同一文本进行编码,邀请专家匿名评价与检验两者的分析质量,发现:生成式AI对人类研究者起着十分重要的辅助与支持作用;人工与机器在质性数据的编码上有相近表现,但在扎根理论、主题分析两类不同的质性方法中呈现出差异化的分析质量。研究进一步探究出现这一现象的原因,提出未来“生成式AI+质性研究”需要加强人机合作与协同、多模态数据智能分析、质性研究全流程介入等。With the emergence and proliferation of generative AI,its involvement in qualitative research has increasingly expanded.The academic community has conducted numerous comparative studies on the practical applications of human and machine involvement in qualitative research,demonstrating the positive role of generative AI in qualitative studies.This study aims to conduct a comparative analysis of analytical quality between human and machine in qualitative research(grounded theory and thematic analysis)within the Chinese context,utilizing China’s domestically developed generative AI tools and family education interview texts as the data foundation,thereby enriching domestic research in this thematic area.The research employs parallel human and machine coding of identical texts,with anonymous expert evaluations assessing the analytical quality of both approaches.Findings reveal that generative AI plays a significant supportive role in assisting human researchers,with comparable performance between human and machine in qualitative data coding.However,distinct analytical quality differences emerge between human and machine across the two qualitative methodologies(grounded theory vs.thematic analysis).This study further investigates the underlying reasons for these variations while proposing future directions for"AI+qualitative research,"including human-machine collaboration,intelligent analysis of multimodal data,and AI integration throughout the entire qualitative research workflow.

关 键 词：生成式 AI 人工智能 人机合作 质性研究 分析质量 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]