Generative AI Policy

The Journal of Diabetes, Metabolism and Endocrinology recognizes that generative artificial intelligence (AI) and AI-assisted technologies are increasingly used in research, manuscript preparation, peer review, data analysis, and scholarly communication. The journal supports responsible use of these technologies while maintaining human accountability, research integrity, transparency, confidentiality, and scientific accuracy.

This policy is informed by relevant principles and recommendations from the International Committee of Medical Journal Editors (ICMJE) and the Committee on Publication Ethics (COPE).

1. General Principle

Generative AI tools may be used as an assistive technology, provided that their use does not compromise the integrity, originality, confidentiality, or accuracy of scholarly work.

AI-generated or AI-assisted content must always be appropriately reviewed and verified by human authors or editors.

AI systems cannot assume responsibility for published research and therefore cannot be listed as authors or co-authors.

2. Author Responsibility

Authors remain fully responsible for the content of their manuscript, regardless of whether AI tools were used during its preparation.

Authors must:

  • Verify the accuracy of AI-assisted content.
  • Check all references and citations.
  • Correct factual or methodological errors.
  • Ensure that AI-generated text does not contain plagiarism or inappropriate text reuse.
  • Protect confidential and personal information.
  • Ensure that the final manuscript accurately represents the underlying research.
  • Disclose material use of generative AI where required by this policy.

Use of an AI tool does not transfer responsibility for the manuscript from the authors to the technology provider.

3. Acceptable Uses of Generative AI

Subject to appropriate verification and disclosure, authors may use generative AI tools for limited assistance such as:

  • Language improvement.
  • Grammar and spelling correction.
  • Improving readability.
  • Translation assistance.
  • Formatting assistance.
  • Brainstorming or organizational support.
  • Assistance with coding or computational workflows.
  • Administrative manuscript preparation.

Authors remain responsible for verifying the resulting content.

4. Substantive Use of Generative AI

Where generative AI materially contributes to the intellectual or textual development of a manuscript, authors should disclose its use.

Examples may include AI assistance with:

  • Generating substantial portions of manuscript text.
  • Developing or substantially modifying analytical code.
  • Producing substantial methodological content.
  • Generating or modifying figures.
  • Conducting substantial data analysis.
  • Summarizing or interpreting research material.

The disclosure should identify, where relevant:

  • The AI tool or system used.
  • The purpose for which it was used.
  • The extent of its contribution.
  • The relevant stage of the research or manuscript preparation.

5. AI and Authorship

Generative AI tools must not be listed as authors, co-authors, contributors with authorship status, or corresponding authors.

AI systems cannot:

  • Accept responsibility for the research.
  • Approve the final manuscript.
  • Declare conflicts of interest.
  • Provide informed consent.
  • Respond to research-integrity investigations.
  • Take legal or ethical responsibility for published work.

Human authors remain accountable for all published content.

6. AI-Generated Text

Authors must carefully review AI-generated text before incorporating it into a manuscript.

AI systems may generate:

  • Incorrect facts.
  • Fabricated references.
  • Misleading interpretations.
  • Unsupported scientific claims.
  • Inappropriate citations.
  • Biased or incomplete explanations.

Authors must independently verify the accuracy and scientific validity of AI-assisted material.

7. AI-Generated References and Citations

Authors are responsible for verifying every citation and reference generated or suggested by an AI system.

References must correspond to genuine and relevant scholarly sources.

Fabricated, inaccurate, irrelevant, or unverifiable references are unacceptable, regardless of whether they were generated by an AI tool.

Authors should not cite an AI system as a substitute for the original scholarly source when the original source is available.

8. AI and Research Data

Authors must not use generative AI tools in ways that compromise the confidentiality, security, or integrity of research data.

Particular caution is required for:

  • Patient information.
  • Medical records.
  • Personally identifiable information.
  • Protected health information.
  • Confidential datasets.
  • Unpublished research data.
  • Proprietary information.
  • Restricted institutional information.

Authors should not upload confidential or identifiable research information to external AI services unless appropriate authorization, security, privacy, and data-protection safeguards are in place.

9. AI in Medical and Clinical Research

Because the journal publishes research related to medicine, diabetes, metabolism, and endocrinology, authors must exercise particular caution when AI tools are used with clinical or patient-related information.

AI assistance must not result in:

  • Disclosure of patient identity.
  • Unauthorized processing of medical information.
  • Fabrication of clinical findings.
  • Alteration of patient data.
  • Misrepresentation of clinical outcomes.
  • Unsupported medical recommendations.

AI-generated clinical interpretations must be independently evaluated by appropriately qualified researchers or healthcare professionals.

10. AI-Generated Images, Figures, and Illustrations

Authors must disclose the use of generative AI to create or substantially modify images, figures, illustrations, or other visual material where such use is permitted.

AI-generated visual content must not:

  • Falsify research findings.
  • Alter experimental results deceptively.
  • Add, remove, or manipulate scientific evidence in a misleading manner.
  • Misrepresent clinical or experimental observations.

The use of AI to manipulate research images or scientific evidence in a manner that changes the underlying findings may constitute research misconduct.

11. AI-Assisted Data Analysis and Coding

AI tools may assist researchers with coding, statistical workflows, or computational analysis.

However, authors remain responsible for:

  • Validating the code.
  • Checking calculations.
  • Verifying statistical methods.
  • Confirming the reproducibility of the analysis.
  • Ensuring that AI-generated code does not introduce methodological errors.

Where AI makes a substantive contribution to computational analysis, appropriate disclosure is encouraged.

12. Generative AI in Peer Review

Reviewers must maintain strict confidentiality when using AI-assisted technologies.

Reviewers should not upload confidential manuscripts, unpublished research, reviewer reports, or other protected editorial information into public or external generative AI systems where confidentiality cannot be assured.

Reviewers remain responsible for the content and recommendations of their review.

AI tools must not replace the reviewer's independent scholarly judgment.

13. Generative AI and Editorial Decision-Making

Editors may use appropriate AI-assisted tools for limited administrative or editorial purposes where confidentiality and data-protection requirements can be maintained.

AI systems must not independently make final decisions concerning:

  • Acceptance.
  • Rejection.
  • Retraction.
  • Ethical misconduct.
  • Authorship disputes.
  • Conflicts of interest.

Final editorial decisions must remain under the responsibility of qualified human editors.

14. AI and Publication Ethics Investigations

AI-generated content may be considered as part of an editorial assessment, but AI-detection software or automated AI classifiers should not be treated as definitive evidence of misconduct on their own.

AI-detection results may produce false positives or false negatives.

Where concerns arise, the journal may consider:

  • Manuscript content.
  • Author explanations.
  • Research records.
  • Data and supporting documentation.
  • Citation accuracy.
  • Other evidence relevant to the matter.

15. Disclosure Statement

Where disclosure is required, authors should include an appropriate statement in the manuscript.

For example:

“The authors used [name of AI tool] for language and/or editorial assistance during manuscript preparation. The authors reviewed and verified all AI-assisted content and remain fully responsible for the accuracy, originality, and integrity of the published work.”

The disclosure should accurately describe the actual use of AI and should not claim that AI was used when it was not.

16. AI and Plagiarism

AI-generated text does not automatically constitute plagiarism; however, authors remain responsible for ensuring originality and appropriate attribution.

Authors must not use generative AI to:

  • Paraphrase copyrighted material deceptively.
  • Conceal plagiarism.
  • Circumvent similarity screening.
  • Generate fabricated citations.
  • Misrepresent another person's ideas as their own.

The journal may conduct plagiarism and similarity screening using appropriate tools.

17. AI and Copyright

Authors are responsible for ensuring that AI-assisted content complies with applicable copyright, licensing, intellectual-property, and attribution requirements.

Where third-party material is incorporated into AI-assisted output, authors must ensure that appropriate rights and permissions exist where required.

18. Undisclosed or Misleading AI Use

Failure to disclose substantive AI use where disclosure is required may constitute a publication-ethics concern.

If significant undisclosed AI use is identified, the journal may request clarification or take appropriate editorial action depending on the circumstances.

Potential actions may include:

  • Request for disclosure.
  • Manuscript revision.
  • Additional editorial review.
  • Rejection.
  • Correction of the published record.
  • Expression of Concern.
  • Retraction where AI use materially compromises the integrity of the publication.

19. Compliance with COPE and ICMJE

The journal follows applicable COPE principles and guidance concerning AI-assisted scholarly publishing, confidentiality, authorship, peer review, and research integrity.

The journal also follows relevant ICMJE Recommendations regarding the use of AI and machine-learning tools in scholarly publishing, including the principles that authors remain responsible for the accuracy and integrity of their work and that AI tools cannot satisfy authorship requirements.

20. Right to Request Additional Information

The Editorial Office may request information concerning the use of AI where it is relevant to:

  • Research integrity.
  • Authorship.
  • Data processing.
  • Manuscript preparation.
  • Image generation or modification.
  • Statistical analysis.
  • Publication ethics.

Authors are expected to cooperate with reasonable editorial requests.

21. Final Principle

The Journal of Diabetes, Metabolism and Endocrinology supports responsible innovation in scholarly publishing but places human accountability above automated generation.

Generative AI may assist researchers, authors, reviewers, and editors, but it cannot replace qualified human judgment or responsibility. The use of AI must never compromise research integrity, patient confidentiality, scientific accuracy, originality, transparency, or ethical standards.

The journal reserves the right to update this policy as generative AI technologies, ethical standards, regulatory requirements, and international publishing guidance continue to evolve.