General Principles

AI and automated tools may support legitimate research, writing, analysis, editing, and publishing activities, but they must not replace the intellectual responsibility, critical judgment, verification, or accountability of human participants in the scholarly publishing process.

Use of AI by Authors

Authors may use AI or AI-assisted technologies where their use is appropriate to the research or preparation of the manuscript, provided that scholarly responsibility remains with the authors.

Authors must critically review and verify any AI-assisted output and remain responsible for the accuracy, originality, validity, integrity, citations, data, interpretations, and conclusions presented in the manuscript.

• AI tools must not replace the author's independent scholarly judgment or interpretation of research findings.
• Authors must verify factual statements, calculations, references, quotations, data analysis, and other outputs generated or suggested by AI systems.
• Authors must ensure that AI-assisted content does not introduce fabricated information, false references, inappropriate attribution, misleading claims, or undisclosed manipulation.
• AI-assisted research must comply with applicable requirements concerning research ethics, privacy, confidentiality, intellectual property, data protection, and informed consent.
• Authors remain responsible for identifying and mitigating material limitations or biases introduced by AI-based methods where these may affect the research.

Disclosure of Generative AI Use

Authors should disclose the use of generative AI or similar automated content-generation tools when such use goes beyond straightforward spelling correction, grammar correction, language editing, formatting, or other basic editorial assistance.

Material AI use may include, for example:

• generating or substantially rewriting manuscript text;
• generating summaries, interpretations, or research arguments;
• assisting materially with coding, computational analysis, modelling, or algorithm development;
• generating or materially modifying figures, images, illustrations, tables, or other research content; or
• other AI-assisted activities that materially contributed to the research or manuscript.

AI and Authorship

Artificial intelligence systems, large language models, chatbots, and other automated tools cannot qualify as authors or co-authors.

Authorship requires human accountability for the integrity of the work, approval of the final manuscript, responsibility for scholarly claims, and the ability to respond to questions about the research. AI systems cannot fulfil these responsibilities.

AI-Generated Content and References

Generative AI systems should not be treated as authoritative scholarly sources and should not be cited as substitutes for original literature or primary evidence.

Authors are responsible for locating, verifying, and citing the actual scholarly or primary sources supporting statements made in the manuscript.

References suggested by AI must be independently verified. Fabricated, incorrect, incomplete, or unverifiable references are unacceptable.

AI as Part of the Research Method

When AI, machine learning, deep learning, large language models, computer vision, automated decision systems, or similar technologies constitute part of the research methodology, their use should be described with sufficient methodological transparency for scholarly evaluation.

Where relevant, authors should report information such as the model or system used, version or configuration, data source, preprocessing, validation approach, relevant prompts or procedures, evaluation metrics, and limitations.

The appropriate level of reporting depends on the research design and should be sufficient to allow readers to understand, evaluate, and where reasonably possible reproduce or verify the reported method.

Use of Generative AI by Reviewers

Reviewers must protect the confidentiality of manuscripts and peer-review materials.

Reviewers should not upload submitted manuscripts, unpublished data, figures, supplementary materials, author information, or confidential editorial communications to generative AI systems where doing so could disclose confidential or proprietary information.

Generative AI should not be used to produce the substantive scientific assessment, recommendation, or peer-review report on behalf of the reviewer.

Limited language editing or rewriting assistance may be used only where confidentiality is protected and the reviewer remains fully responsible for the final review. Any material AI-assisted preparation of reviewer comments should be disclosed to the journal where appropriate.

Use of Generative AI by Editors

Editors must preserve manuscript confidentiality and should not upload unpublished manuscript content or confidential peer-review information to external generative AI systems where confidentiality cannot be assured.

Generative AI must not independently determine whether a manuscript should be accepted or rejected and must not replace human editorial judgment.

Editors remain responsible for all editorial decisions and for evaluating the reliability and appropriateness of any automated assistance used during manuscript processing.

Automated Tools Used by the Journal

DEMETER may use automated or AI-assisted tools to support editorial integrity checks, manuscript administration, similarity screening, or related editorial functions.

Where automated tools are used to identify potential integrity concerns, their results are treated as indicators requiring human editorial verification rather than as automatic evidence of misconduct.

Editors must consider the context, limitations, possible false positives, and reliability of automated results before taking editorial action.

AI-Generated Content Detection

DEMETER does not determine compliance with this policy solely through an AI-generated content percentage or automated detection score.

AI-detection systems may be used as supporting editorial tools where appropriate, but their output must be reviewed by a human editor and interpreted in context.

A numerical AI-detection score alone is not sufficient evidence that an author has violated this policy and will not, by itself, determine acceptance, rejection, or allegations of misconduct.

Undisclosed or Inappropriate AI Use

When the editorial team identifies potentially material, undisclosed, misleading, or inappropriate use of AI, authors may be asked to clarify how the technology was used and to revise or provide supporting information where necessary.

Concerns involving fabricated content, falsified data, false citations, plagiarism, privacy violations, authorship misrepresentation, or other research integrity issues are handled under the journal's applicable Publication Ethics, Plagiarism Policy, or Correction & Retraction Policies.