JMIR Publications Editorial Policy on the use of ChatGPT and similar tools in the preparation of manuscripts
The full policy is published in this editorial on "Best Practices for Using AI Tools as an Author, Peer Reviewer, or Editor" in the Journal of Medical Internet Research. Authors are required to adhere to the policies with regards to author accountability, transparency, and maintenance of confidentiality of any content generated by AI and incorporated into a submitted manuscript for peer review and potential publication in any JMIR Publications journal.
Because of the rapidly evolving nature of AI technologies, related policies, regulation, and best practices, JMIR Publications may periodically modify this editorial policy. It is the responsibility of the authors to adhere to these principles and practices provided.
**Authors are advised to carefully read the full published editorial policy on generative AI usage to ensure their adherence to the required appropriate statements and/or supplementary materials.**
In brief, practical important author considerations with respect to generative AI usage in generating manuscript content also includes:
- If ChatGPT generated text is used, it should be presented as blockquote, textbox or figure, making clear that this text comes from ChatGPT.
- ChatGPT should never be listed as author [2]
- We do allow the use of ChatGPT in ideation, but ask for full disclosure on the extent of use of ChatGPT in the ideation or writing process.
- Disclosure: We recommend to do so within the methods or acknowledgements sections; we also ask to keep the complete original ChatGPT transcripts on file and submit all ChatGPT conversations used in the preparation of the manuscript as a Multimedia Appendix so reviewers and readers can assess the extent of text used from ChatGPT
- We warn against taking statements or references from ChatGPT at face value - ChatGPT sometimes strings words together without understanding the meaning behind it, and we have seen it making up references that don't exist [3], so if references are suggested by ChatGPT, authors are responsible to check and read them and make sure that they say what ChatGPT purports they say. Authors must fact-check statements made by ChatGPT and are ultimately responsible for the content of their article.
- The submission of manuscripts that are solely or to a large extent written by ChatGPT without proper disclosure constitutes academic misconduct
References:
1. Leung T, de Azevedo Cardoso T, Mavragani A, Eysenbach G. Best Practices for Using AI Tools as an Author, Peer Reviewer, or Editor. J Med Internet Res 2023;25:e51584. https://www.jmir.org/2023/1/e51584
DOI: 10.2196/51584
2. COPE position statement on Authorship and AI
3. Eysenbach G. The Role of ChatGPT, Generative Language Models and Generative Artificial Intelligence in Medical Education: An Conversation with ChatGPT and a Call for Papers (Editorial). JMIR Med Educ 2023;9:e46885 doi: https://doi.org/10.2196/46885
(V2023-August 31, 2023 - This policy is subject to change)