出版科学 ›› 2024, Vol. 32 ›› Issue (06): 15-25.

• 理论探索 • 上一篇    

国际出版机构使用人工智能工具的出版伦理政策分析

  

  1. 四川大学华西医院
  • 收稿日期:2024-02-20 修回日期:2024-09-04 出版日期:2024-11-26 发布日期:2024-11-26
  • 作者简介:董敏,四川大学华西医院编辑; 刘雪梅(通讯作者),四川大学华西医院编审; 杜亮,四川大学华西医院教授、博士生导师; 雷芳,理学博士,四川大学华西医院编辑。

Analysis on Publishing Ethics Policies of International Publishers for the Use of Artiffcial Intelligence Tools

  1. West China Hospital, Sichuan University
  • Received:2024-02-20 Revised:2024-09-04 Online:2024-11-26 Published:2024-11-26

摘要: 从科学网(Web of Science)的期刊引证报告(Journal Citation Reports)数据库排名前 70 的出版机构中筛选出 23 家,分析其人工智能(AI)出版伦理政策。在作者使用 AI 方面, 出版机构主要从署名、AI 使用边界、披露形式进行了规范。出版机构对 AI 身份问题已达成共识, 即 AI 不能成为署名作者;出版机构允许在研究和写作中使用 AI,但使用者需要对内容负责且充分披露,而目前只有少数(26.1%)机构明确规定了作者的 AI 使用边界,且各机构之间对于如何披露的规定存在差异。关于审稿者使用 AI 的政策,鉴于保密性要求,机构一般不允许审稿者将稿件上传至无法保证安全性的 AI 工具,但关于如何披露和违反 AI 使用政策的惩罚措施还有待完善。针对上述问题,建议出版机构从伦理、技术和法律三个层面,以及协同治理的角度,与学术共同体和管理部门合作,共同制定学术出版中 AI 使用的伦理规范,以营造健康的学术生态环境。

Abstract:

The article selectes 23 publishing institutions from the top 70 in the Journal Citation Reports database of Web of Science and analyzes their artificial intelligence(AI)publishing ethics policies. In terms of the author’ s use of AI,publishing institutions have mainly standardized the authorship, AI usage boundaries and disclosure forms. Publishing institutions have reached a consensus on the issue of AI identity. That is, AI cannot become a coauthor. Publishing institutions allow the use of AI in research and writing, but require authors’ responsibility and full disclosure. Currently only a few institutions(26.1%) have clearly defined the boundaries of authors’ AI usage,and there are differences among institutions on how to disclose them. Regarding the policy on the use of AI by reviewers,due to confidentiality requirements, institutions generally do not allow reviewers to upload manuscripts to AI tools that cannot guarantee security. However, there is still room for improvement on how to disclose and punish violations of AI usage policies. In response to the above issues, it is suggested that publishing institutions collaborate with academic communities and management departments from the perspectives of ethics, technology and law, as well as collaborative governance, to jointly develop ethical standards for AI applications in academic publishing, in order to create a healthy academic ecosystem.

中图分类号: