AI POLICIES

Green Medical Journal (GMJ) recognizes that artificial intelligence (AI), including generative AI and AI-assisted technologies, may support scientific writing, research, data analysis, and editorial work in medicine and health sciences. However, the use of these technologies must not compromise research integrity, patient confidentiality, scientific validity, research ethics, or the reliability of medical evidence.

This policy is intended to ensure the responsible and transparent use of AI in manuscripts submitted to or published by GMJ. It applies to research involving human participants, medical records, laboratory investigations, animal experiments, clinical and biomedical images, public health, and other areas within the journal’s scope.

This policy is informed by relevant guidance from the Committee on Publication Ethics (COPE), the World Association of Medical Editors (WAME), and the International Committee of Medical Journal Editors (ICMJE).

1. Scope and Definitions

This policy applies to authors, reviewers, editors, and other individuals involved in the preparation, assessment, and publication of manuscripts in GMJ.

For this policy:

Artificial Intelligence (AI) refers to computational systems capable of performing tasks such as generating text, summarizing information, translating language, analyzing data, creating images, or assisting with scientific interpretation.
Generative AI refers to AI systems that produce new content, including text, images, code, tables, and other materials.
AI-assisted research refers to the use of AI as part of a research method, such as medical image analysis, data classification, prediction modeling, or other analytical procedures.
AI-assisted writing refers to the use of AI to support manuscript preparation, including drafting, rewriting, translation, summarization, or language improvement.

The use of AI must be appropriate to the research purpose and must not replace the scientific judgment and accountability of qualified human researchers.

2. Authorship and Accountability

AI tools cannot be listed as authors or co-authors of manuscripts submitted to GMJ because they cannot assume responsibility for the accuracy, integrity, originality, and ethical conduct of the work.

Authors are responsible for:

Ensuring that all reported research findings are accurate and supported by the original data.
Verifying the accuracy and relevance of references, citations, scientific statements, and medical information generated or suggested by AI.
Ensuring that the manuscript represents their own scholarly work and meets applicable ethical requirements.
Reviewing AI-generated content for errors, bias, fabricated information, unsupported conclusions, and misleading medical claims.
Taking full responsibility for the content of the submitted manuscript, including any material produced with AI assistance.

AI-generated medical information must not be accepted without appropriate verification against reliable scientific evidence and the study’s actual findings.

3. Acceptable Use of AI by Authors

Authors may use AI tools for appropriate purposes, provided that the use does not compromise research integrity or confidentiality. Acceptable uses may include:

Improving grammar, spelling, readability, and academic language.
Translating or refining text, provided that the scientific meaning is preserved.
Assisting with manuscript organization, outlines, or summaries.
Supporting literature exploration, provided that authors independently verify the sources and references.
Assisting with programming, statistical workflows, data organization, or preliminary analysis, provided that methods and results are independently checked.
Conducting AI-assisted analysis as part of a defined research method, such as medical image processing or predictive modeling, when scientifically justified and transparently reported.

For research involving patients, medical records, laboratory findings, animal experiments, or clinical images, authors must ensure that AI-assisted procedures are appropriate, reproducible where applicable, and consistent with research ethics and data protection requirements.

AI must not be used as a substitute for professional medical judgment, independent statistical review when needed, or appropriate scientific interpretation.

4. Unacceptable Use of AI by Authors

The following practices are prohibited:

Fabrication or falsification of data: Generating, altering, or inventing patient records, laboratory results, experimental observations, statistical findings, or other research data and presenting them as genuine observations.
Fabricated references: Citing nonexistent publications, invented authors, false DOIs, or references that do not support the stated claims.
Misrepresentation of research findings: Using AI-generated explanations or conclusions that are inconsistent with the actual data or are presented as established medical evidence without adequate support.
Plagiarism and misappropriation: Using AI to disguise copied material, misrepresent the originality of a manuscript, or reproduce another person’s work without appropriate attribution.
Manipulation of scientific images: Creating, removing, adding, or modifying elements in clinical photographs, radiological images, histopathological images, microscopic images, or laboratory evidence in ways that misrepresent the original findings.
Violation of confidentiality: Uploading identifiable patient information, confidential medical records, unpublished research data, or other protected information to public or unauthorized AI systems.
Misleading clinical claims: Using AI-generated content to assert diagnostic accuracy, treatment effectiveness, safety, prognosis, or clinical recommendations that are not supported by the study design and results.
Circumvention of ethical requirements: Using AI to conceal the absence of ethics approval, informed consent, animal research approval, or other required safeguards.

AI-generated synthetic data may only be used when it is explicitly part of a legitimate research design, clearly identified as synthetic, and reported transparently. Synthetic data must not be represented as observations obtained from real patients, participants, or experimental animals.

5. Disclosure of Generative AI Use

Authors must disclose the use of generative AI when it contributes substantively to manuscript preparation, scientific content, data analysis, interpretation, or the creation of figures and other research materials.

Routine spelling, grammar, and formatting assistance that does not generate substantive content generally does not require a separate declaration, unless otherwise requested by GMJ.

The disclosure should identify:

The name of the AI tool and, where available, the model or version used.
The purpose of using the tool.
The relevant section or component of the manuscript affected.
The steps taken by the authors to verify the generated content or output.

The disclosure should be placed in an AI Use Declaration section before the References, unless the use is more appropriately described in the Methods section. If AI forms part of the research methodology, the authors must also describe its role, procedures, and relevant validation in the Methods section.

Example declaration:

AI Use Declaration: The authors used [name of tool and model/version, if available] to assist with [specific purpose]. The generated content was reviewed and verified by the authors, who take full responsibility for the accuracy, originality, and integrity of the manuscript.

For AI used in research analysis, authors should additionally report the relevant technical details, such as the AI model or software, version, intended function, input data, analytical procedures, and validation methods, as applicable to the study.

If no generative AI was used in the preparation or conduct of the work, authors may state:

AI Use Declaration: No generative AI tools were used in the preparation of this manuscript or in the conduct of the research, except for routine spelling, grammar, or formatting assistance, if applicable.

Authors must provide further information about AI use if requested by the editors.

6. Patient Confidentiality, Clinical Data, and Research Ethics

GMJ requires strict protection of patient confidentiality and research participants’ rights.

Authors must not submit identifiable patient information, medical records, patient photographs, genetic information, or confidential clinical data to public or unauthorized AI services. Removing names alone may not sufficiently protect identity; authors must also consider the risk of re-identification.

Research involving human participants or patient records must comply with applicable ethical approval, informed consent, privacy, and data protection requirements. The use of AI does not remove or replace these obligations.

For animal experiments, including studies evaluating treatments or biological effects in laboratory animals, authors remain responsible for compliance with applicable animal research ethics requirements and for the accurate reporting of experimental procedures and findings.

When AI supports the analysis of clinical, laboratory, radiological, histopathological, or other biomedical data, authors must ensure that:

The analytical approach is appropriate for the research question.
The output is checked against the original data and validated as appropriate.
Limitations, potential bias, and uncertainty are acknowledged.
Claims about diagnosis, treatment, prognosis, or clinical applicability are supported by the available evidence.
AI-generated interpretations are not presented as independently confirmed medical findings without adequate verification.

7. AI-Generated Images, Figures, and Biomedical Evidence

AI-generated images and AI-assisted image processing require particular care in medical and biomedical research.

AI must not be used to fabricate or alter clinical photographs, radiological findings, histopathological specimens, microscopic fields, laboratory images, or experimental results in a way that changes or misrepresents the evidence.

If AI is used for legitimate image analysis, segmentation, enhancement, classification, or other research procedures, authors must describe the relevant methods and disclose the use where it materially affects the research. Original source images and appropriate records of processing must be retained and made available to the journal when requested.

AI-generated illustrations may be considered when they are relevant to the manuscript, permitted by the applicable publishing requirements, and clearly distinguished from actual clinical or experimental evidence. Authors must not present illustrative or synthetic images as photographs or findings obtained from real patients, research participants, or laboratory experiments.

GMJ may request original images, underlying data, or further methodological information when necessary to assess the authenticity and reliability of submitted material.

8. Responsibilities of Peer Reviewers

Peer reviewers must maintain the confidentiality of manuscripts and all associated materials.

Reviewers must not upload submitted manuscripts, unpublished findings, patient information, or confidential supplementary materials to public or unauthorized AI systems. This restriction applies to AI-assisted summarization, critical appraisal, language editing, and other forms of manuscript assessment where confidentiality cannot be assured.

Reviewers are expected to:

Evaluate the scientific quality, originality, methods, statistical analysis, ethical compliance, and interpretation of the manuscript independently.
Verify any concerns about AI-generated references, claims, images, or research findings before reporting them.
Avoid delegating confidential or substantive peer-review judgments to AI tools.
Inform the editor if they identify credible concerns about possible inappropriate AI use.

Any permitted AI assistance must not compromise confidentiality, reviewer independence, or the quality of the assessment.

9. Responsibilities of Editors

Editors are responsible for ensuring that AI is used in accordance with GMJ’s publication ethics standards.

Editors must:

Protect the confidentiality of submitted manuscripts and related editorial communications.
Not upload confidential submissions to public or unauthorized AI systems.
Make editorial decisions based on human assessment of scientific merit, ethical compliance, and the journal’s publication criteria.
Assess disclosed AI use in the context of the manuscript’s research design and content.
Request clarification, original data, source images, or additional methodological details when concerns arise.
Treat AI-detection results as preliminary indicators rather than conclusive evidence of misconduct.

Editorial decisions must not be based solely on automated AI-detection scores. Any concern must be assessed fairly, with appropriate consideration of the evidence and an opportunity for the authors to respond.

10. Handling Potential Breaches of This Policy

If GMJ identifies possible inappropriate AI use, the journal may request clarification, disclosure details, original data, source images, research records, or other supporting documentation from the authors.

Depending on the seriousness of the concern and the available evidence, GMJ may take one or more of the following actions:

Request a correction or clarification of the manuscript.
Require additional disclosure or methodological information.
Request revisions or supporting evidence before further consideration.
Reject a manuscript when serious concerns remain unresolved.
Publish a correction, expression of concern, or retraction for a published article when justified.
Refer serious concerns to the relevant institution or other appropriate authority when necessary.

Allegations will be assessed in accordance with applicable publication ethics guidance and GMJ procedures. The use of an AI-detection tool alone is not sufficient to establish misconduct.

11. Review and Updates of This Policy

GMJ may periodically review and revise this policy in response to developments in AI technology, medical research practices, publication ethics, and relevant recommendations from COPE, WAME, ICMJE, or other appropriate organizations.

The journal will communicate material changes through its official website.

Effective date: To be determined by the GMJ editorial board following approval.

References

Committee on Publication Ethics (COPE). Authorship and AI tools. https://publicationethics.org/guidance/cope-position/authorship-and-ai-tools
World Association of Medical Editors (WAME). Chatbots, generative AI, and scholarly manuscripts. https://wame.org/page3.php?id=106
International Committee of Medical Journal Editors (ICMJE). Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals. https://www.icmje.org/recommendations/
International Committee of Medical Journal Editors (ICMJE). AI Use by Authors. https://icmje.org/recommendations/browse/artificial-intelligence/ai-use-by-authors.html