摘要：

目的 探讨图书情报（LIS）领域期刊对生成式人工智能（AI）的政策规范，为推动AI技术在LIS领域学术出版和科学研究的规范化应用提供参考依据。 方法 选取LIS领域的SSCI和ESCI期刊，采用LDA主题建模与Nvivo质性编码相结合的方法，从索引类型、期刊分区与出版商背景等结构性维度以及作者、审稿人与编辑等角色维度展开多维分析。 结果 AIGC政策在索引、分区和出版商等维度呈现差异化；作者政策关注写作过程、数据分析等，而审稿/编辑政策强调保密性、完整性等；尽管部分期刊已出台政策治理，但仍存在覆盖不均衡、差异层级化及问题边界不清晰等问题。 结论 期刊AIGC政策尚未形成完备的治理体系，应构建“顶层标准引领-中层平台支撑-基层贯彻落实”的体系，助力学术出版在智能时代实现技术赋能与科研诚信的协同共进。

关键词: 生成式人工智能, 学术出版政策, 图书情报学, 科研诚信

Abstract:

Purposes This study investigates the policy frameworks of library and information science （LIS） journals regarding the use of generative artificial intelligence （AIGC）， aiming to provide insights for the standardized application of AI technologies in LIS scholarly publishing and scientific research. Methods A multi-dimensional analysis was conducted on SSCI- and ESCI-indexed LIS journals using a mixed-methods approach combining LDA topic modeling and Nvivo-based qualitative coding， examining structural dimensions （indexing status， journal quartile， publisher background） and role-specific dimensions （authors， reviewers， and editors）. Findings AIGC policies exhibit significant differentiation across indexing databases， journal rankings， and publishers. Author-related policies primarily focus on writing assistance and data analysis， while policies for reviewers and editors emphasize confidentiality， integrity， and accountability. Although some journals have established governance measures， challenges persist， including uneven policy coverage， hierarchical disparities， and ambiguous boundaries of acceptable use. Conclusions Current AIGC policies in LIS journals lack a comprehensive governance system. A tiered framework of “top level standard setting-mid level platform support-grassroots implementation and enforcement” should be established to promote the synergistic advancement of technological empowerment and research integrity in scholarly publishing in the AI era.

Key words: Generative artificial intelligence, Scholarly publishing policy, Library and information science, Research integrity