Artificial Intelligence (AI) Use Policy
The Indo Eurasian Journal of Medical Sciences (IEJMS) recognises the rapid advancement of artificial intelligence, large language models and machine-learning tools in scientific research. To maintain the highest standards of scientific integrity, patient safety and transparency, IEJMS has established strict boundaries regarding the use of generative AI in submitted manuscripts.
This policy applies to all authors, peer reviewers and editorial staff.
1. AI and authorship
AI cannot be an author
- No legal standing. Artificial intelligence tools cannot be listed as authors on any submitted manuscript.
- No accountability. AI tools cannot fulfil the ICMJE criteria for authorship because they cannot understand conflicts of interest, hold copyright, or take legal and moral responsibility for the accuracy and integrity of the published work.
2. Acceptable and unacceptable uses of generative AI
Acceptable uses (disclosure required)
- Language and formatting. Authors may use generative AI tools strictly to improve the readability, grammar or translation of their human-authored text.
- Literature search. AI-assisted search engines may be used to identify relevant literature, provided the authors manually verify all identified citations.
Strictly prohibited uses
- Data generation (fabrication). AI must never be used to generate, simulate or alter raw research data, clinical datasets or patient demographics. This constitutes scientific fraud.
- Image generation and manipulation. The use of generative AI to create or alter clinical images, radiological scans, micrographs or graphical data plots is strictly forbidden.
- Clinical assertions. AI must not be used to generate medical diagnoses, clinical conclusions or unverified scientific claims within the discussion or conclusion sections.
3. Mandatory disclosure requirements
If authors use generative AI tools to assist in the writing process, they must explicitly declare this use.
- Location. The disclosure must be placed in a dedicated section titled “AI Use Disclosure” immediately following the required declarations, before the references.
- Format. The declaration must explicitly state the name of the tool, the version used, the manufacturer, and exactly which parts of the manuscript were processed.
Example statement. “During the preparation of this work, the authors used [name of tool / version] in order to [improve English language grammar and readability]. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.”
Null declaration. If no AI tools were used, authors must state: “The authors declare that no artificial intelligence (AI) or generative AI tools were used in the preparation of this manuscript.”
4. Human accountability and liability
- Ultimate responsibility. The human authors are entirely and solely responsible for all content within the manuscript, including any text generated or modified by AI.
- Hallucinations. AI models are known to generate highly plausible but entirely fabricated information. Authors must manually verify every data point, clinical claim and citation.
- Plagiarism. Generative AI tools may inadvertently reproduce copyrighted text without proper attribution. Authors are liable for any plagiarism or copyright infringement introduced by an AI tool.
5. AI in medical data science and methodologies
This policy governs generative AI used in manuscript preparation, not AI used as a core scientific methodology — for example, training a machine-learning model to detect tumours in MRI scans.
- Methodological disclosure. If AI, machine learning or deep learning algorithms are the subject of the research, the methodology must be exhaustively detailed in the Materials and Methods section.
- Reporting standards. Studies developing or validating AI models for medical diagnosis or prognosis must strictly adhere to TRIPOD+AI (for prediction models) or CLAIM (for medical imaging) reporting guidelines. Code and training datasets should be made publicly available via recognised repositories.
6. AI use by peer reviewers and editors
- Confidentiality breach. Peer reviewers and handling editors are strictly prohibited from uploading any portion of an unpublished manuscript — including abstracts, text, data or figures — into public or proprietary generative AI tools.
- Data security. Uploading confidential medical research into third-party language models violates the authors' copyright and breaches the journal's confidentiality agreements.
- Human judgement. Peer review reports must be the original, critical work of the invited clinical or scientific expert.