期刊AI政策与信息资源管理研究AI4SS范式的协调研究OACHSSCD

[Purpose/Significance]Generative Artificial Intelligence(GenAI)has rapidly penetrated all stages of social science research,catalyzing rise of the Artificial Intelligence for Social Science(AI4SS)paradigm and reshaping the landscape of the research.This transformation spans research tools,synthetic data,simulation experiments,and human-AI collaboration.Against this backdrop,the global scholarly publishing sector faces an escalating governance dilemma:conventional academic regulatory systems,designed for traditional research paradigms,fail to accommodate the emerging methodologies,ethical risks,and uncertainties introduced by GenAI.Current journal AI policies tend to adopt strict regulatory measures,which go against the original intention of AI4SS to promote innovation.Existing research on journal AI policies primarily concentrates on descriptive summarization of individual journals'policy provisions,and most relevant recommendations are proposed solely from the perspective of ethical norms.Therefore,studying journal AI policies from the perspective of proactively embracing AI4SS research paradigm is of great significance.[Method/Process]First,the study collected original journal AI policies from two major data sources:official websites and social media platforms,such as WeChat Official Accounts.Second,the study examined these policies from multiple perspectives,including the organization and retrieval of journal AI policies,the common content of journal AI policies,the disciplinary charac-teristics of journal AI policies,the adoption of general AI guidelines,and the dynamic evolution of journal AI policies.Third,the study compared the positional conflicts between journals and AI4SS with respect to AI application scenarios and normative requirements.[Result/Conclusion]The findings reveal that current journals'AI policies have multiple issues,including insufficient transparency and standardization in policy organization,a misaligned trade-off between normative clarity and operational flexibility,unscientific and incomplete AI content detection indicator systems,excessively rigid and one-sided constraints on peer reviewers,ambiguous attitudes towards synthetic data,and ill-defined boundaries for human-AI collabo-rative research.In response to these limitations,and in light of the characteristics of research under AI4SS paradigm and the position of journals,the study proposes a"4+1+1"governance framework for journal AI policies and puts forward tar-geted optimization recommendations."4"refers to a general AI policy governance framework com-prising four pillars:explicit and strict regulation,flexible response,proactive guidance,and deferred policy formulation."1+1"means that,based upon this general governance framework,journals continuously make flexible adjustments as time evolves,and every journal AI policy has journal-specific characteristics.The study points out three reasons why GenAI cannot be listed as an author on research papers.Finally,the study categorizes human-AI collaboration under AI4SS paradigm into three types according to the depth of GenAI's involvement in research and its degree of subjectivity in the research workflow,and pro-poses differentiated measures for each type.