Ng Jeremy Y, Krishnamurthy Malvika, Deol Gursimran, Al-Khafaji Wid Al-Zahraa, Balaji Vetrivel, Abebe Magdalene, Adhvaryu Jyot, Karrthik Tejas, Mohanakanthan Pranavee, Vellaparambil Adharva, Bouter Lex M, Haynes R Brian, Iorio Alfonso, Lokker Cynthia, Maisonneuve Hervé, Marušić Ana, Moher David
Centre for Journalology, Ottawa Methods Centre, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
Institute of General Practice and Interprofessional Care, University Hospital Tübingen, Tübingen, Germany.
Res Integr Peer Rev. 2025 Sep 8;10(1):19. doi: 10.1186/s41073-025-00178-8.
Artificial intelligence chatbots (AICs) are designed to mimic human conversations through text or speech, offering both opportunities and challenges in scholarly publishing. While journal policies of AICs are becoming more defined, there is still a limited understanding of how Editors in chief (EiCs) of biomedical journals' view these tools. This survey examined EiCs' attitudes and perceptions, highlighting positive aspects, such as language and grammar support, and concerns regarding setup time, training requirements, and ethical considerations towards the use of AICs in the scholarly publishing process.
A cross-sectional survey was conducted, targeting EiCs of biomedical journals across multiple publishers. Of 3725 journals screened, 3381 eligible emails were identified through web scraping and manual verification. Survey invitations were sent to all identified EiCs. The survey remained open for five weeks, with three follow-up email reminders.
The survey had a response rate of 16.5% (510 total responses) and a completion rate of 87.0%. Most respondents were familiar with AIs (66.7%), however, most had not utilized AICs in their editorial work (83.7%) and many expressed interest in further training (64.4%). EiCs acknowledged benefits such as language and grammar support (70.8%) but expressed mixed attitudes on AIC roles in accelerating peer review. Perceptions included the initial time and resources required for setup (83.7%), training needs (83.9%), and ethical considerations (80.6%).
This study found that EiCs have mixed attitudes toward AICs, with some EICs acknowledging their potential to enhance editorial efficiency, particularly in tasks like language editing, while others expressed concerns about the ethical implications, the time and resources required for implementation, and the need for additional training.
人工智能聊天机器人(AIC)旨在通过文本或语音模仿人类对话,这在学术出版领域既带来了机遇,也带来了挑战。虽然关于AIC的期刊政策越来越明确,但对于生物医学期刊的主编(EiC)如何看待这些工具,人们的了解仍然有限。本次调查研究了EiC的态度和看法,突出了积极方面,如语言和语法支持,以及对在学术出版过程中使用AIC的设置时间、培训要求和伦理考量等方面的担忧。
开展了一项横断面调查,目标是多家出版商旗下生物医学期刊的EiC。在筛选的3725种期刊中,通过网络爬虫和人工核查确定了3381封符合条件的电子邮件。向所有确定的EiC发送了调查邀请。调查持续开放了五周,并发送了三封跟进电子邮件提醒。
该调查的回复率为16.5%(共510份回复),完成率为87.0%。大多数受访者熟悉人工智能(66.7%),然而,大多数人在编辑工作中未使用过AIC(83.7%),许多人表示有兴趣参加进一步培训(64.4%)。EiC认可语言和语法支持等好处(70.8%),但对AIC在加速同行评审中的作用态度不一。看法包括设置所需的初始时间和资源(83.7%)、培训需求(83.9%)以及伦理考量(80.6%)。
本研究发现,EiC对AIC态度不一,一些EiC认可其提高编辑效率的潜力,特别是在语言编辑等任务中,而另一些人则对伦理影响、实施所需的时间和资源以及额外培训的必要性表示担忧。