面向科研场景的生成式AI系统用户使用行为研究进展OACHSSCD

[Purpose/Significance]The rapid advancement of generative artificial intelligence(GenAI)profoundly reshapes knowledge production paradigms in academic research and offers substantial potential to enhance research effi-ciency,stimulate creative problem-solving,and accelerate scholarly workflows.However,the widespread integration of GenAI systems into scientific activities simultaneously introduces complex challenges regarding academic integrity,infor-mation credibility,and ethical governance,which necessitates a comprehensive understanding of how researchers interact with these technologies.Despite the exponential growth in literature addressing GenAI adoption in education and general contexts,a systematic and granular synthesis of empirical evidence specifically concerning researchers'usage behaviors within scientific research scenarios remains notably lacking.This study addresses this critical gap by constructing a holistic research landscape that clarifies the thematic structure,methodological characteristics,and theoretical underpinnings of existing scholarship,and provides actionable references for GenAI tool iteration and future research agenda selection from an information science perspective.[Method/Process]This study adopted a systematic literature review approach aligned with the PRISMA guidelines and conducted a comprehensive search across nine major English and Chinese academic data-bases including Web of Science,Scopus,IEEE Xplore,ACM Digital Library and CNKI.A rigorous dual-reviewer screening process followed predefined inclusion and exclusion criteria and focused on high-quality empirical studies of GenAI usage behaviors in real research contexts.This study identified 73 qualified empirical articles from 11,981 initial records and performed in-depth content analysis with the support of NVivo 12.The analytical framework covered six core dimen-sions:problem types,research objectives,disciplinary directions,content themes,sample characteristics,and research methodologies.The study adopted the Stimulus-Organism-Response(S-O-R)theory as a guiding framework and classi-fied content themes into external stimulus,organism trait,and behavioral response to explain the formation mechanisms of researchers'GenAI usage behaviors.[Result/Conclusion]The findings show that existing research mainly concentrates on impact-oriented,evaluative,and effect-related questions,while strategic and exploratory inquiries stay underrepre-sented.Thematically,the literature forms a complete tripartite framework:external stimulus include technical functiona-lities,task contexts,environmental norms,and social influences;organism traits cover cognitive states,emotional atti-tudes,perceived values,and competency foundations;behavioral responses involve usage patterns,behavioral strategies,and operational characteristics.Methodologically,the field relies heavily on cross-sectional quantitative designs and con-venience sampling,and lacks longitudinal or mixed-methods approaches.The study identifies thematic fragmentation,li-mited research purpose diversity,and insufficient cross-cultural perspectives as major limitations.Several future directions would be proposed:enhancing longitudinal and intervention designs,building dual-dimensional evaluation frameworks,expanding interdisciplinary research,and exploring dynamic interactions among stimulus,organismic mediators,and behavioral outcomes in Information Resource Management contexts.