中国科技期刊研究 ›› 2025, Vol. 36 ›› Issue (7): 835-843. doi: 10.11946/cjstp.202503110240

• AI赋能学术期刊出版专题 • 上一篇    下一篇

人机协同视角下AI辅助审稿的SWOT分析

黄艺聪1,2)()(), 黄谷香2), 孙启艳2), 杨珏2)   

  1. 1) 上海交通大学安泰经济与管理学院,上海市徐汇区华山路1954号 200030
    2) 上海交通大学期刊中心,上海市徐汇区广元西路55号 200030
  • 收稿日期:2025-03-11 修回日期:2025-04-26 出版日期:2025-07-31 发布日期:2025-07-31
  • 作者简介:

    黄艺聪(ORCID:0000-0001-7580-1994),硕士,编辑,E-mail:;

    黄谷香,博士,副编审,编辑部主任;

    孙启艳,硕士,编辑;

    杨 珏,硕士,编辑。

    作者贡献声明: 黄艺聪:论证选题,设计论文框架,撰写及修改全文; 黄谷香:审核论文框架,给出修改建议; 孙启艳:调研与整理文献,参与讨论; 杨 珏:整理文献,参与讨论。
  • 基金资助:
    2024年度上海交通大学期刊中心期刊发展研究基金研究课题(QK-A-2024002)

SWOT analysis of AI-assisted review from a human-machine collaboration perspective

HUANG Yicong1,2)()(), HUANG Guxiang2), SUN Qiyan2), YANG Jue2)   

  1. 1) Antai College of Economics and Management, Shanghai Jiao Tong University, 1954 Huashan Road, Xuhui District, Shanghai 200030, China
    2) Journal Center, Shanghai Jiao Tong University, 55 Guangyuan West Road, Xuhui District, Shanghai 200030, China
  • Received:2025-03-11 Revised:2025-04-26 Online:2025-07-31 Published:2025-07-31

摘要:

【目的】探讨AI在审稿流程中的应用前景,为学术出版数字化、智能化发展提供参考。【方法】梳理近几年的相关文献和国内外审稿政策,运用SWOT分析法系统分析AI辅助审稿的优势、劣势、外部机遇和挑战,构建SWOT分析矩阵。【结果】AI技术在审稿流程中有应用潜力,有助于提高审稿效率和客观性,但需解决幻觉、偏见、安全和信任问题才能实现规模化应用,这有赖于可信的智能审稿系统的开发与监督机制的建立。【结论】未来,若要实现人机协同的AI辅助审稿模式,应提高AI的技术透明性,强化问责制,构建开放算料联盟,并针对AI存在的技术局限性与伦理风险进行规制与治理。通过技术与制度的双重优化,AI有望成为优化学术审稿流程的重要工具。

关键词: AI, 审稿, 生成式人工智能, 学术期刊, 出版伦理

Abstract:

[Purposes] This study explores the application prospects of AI in review process to provide insights for the digital and intelligent development of academic publishing. [Methods] Based on literature research and comparative analyses of review policies in domestic and international academic journals, using SWOT analysis, the study systematically examines the advantages, disadvantages, opportunities, and challenges of AI-assisted review,constructing a SWOT analysis matrix. [Findings] AI technology has potential for application in the review process, which can improve review efficiency and enhance the objectivity of reviews. However, problems such as hallucination, bias, security, and trust need to be addressed to achieve large-scale application. This relies on the development of trustworthy intelligent review systems and the establishment of supervision mechanisms. [Conclusions] In the future, to achieve human-machine collaborative AI-assisted review, it is necessary to enhance AI’s technical transparency, strengthen accountability, establish an open data alliance, and implement regulation and governance for AI's technical limitations and ethical risks. Through the dual optimization of technology and systems, AI has the potential to become a crucial tool in streamlining the academic review process.

Key words: AI, Review, Generative artificial intelligence, Academic journals, Publishing ethics