Plagiarism Screening and Generative AI Policy

All submitted manuscripts will undergo similarity screening before peer review. The journal may use a similarity score of 15% as an internal indicator for closer editorial examination, but the score will not be treated as an automatic measure of plagiarism or as the sole basis for rejection.

Editors will review the nature, extent, location, and sources of the matched text, taking into account properly quoted material, reference lists, standard terminology, methodological descriptions, preprints, theses or dissertations, and the authors’ previously published work. A manuscript may be returned for revision, rejected, or investigated further where the editorial assessment identifies plagiarism, inappropriate text recycling, redundant publication, or another breach of publication ethics. Authors will ordinarily be informed of the concern and given an opportunity to respond where appropriate.

Milkiyah’s policy is informed by relevant COPE guidance on authorship, transparency, accountability, confidentiality, and publication integrity. AI tools, including Large Language Models, cannot be listed as authors because they cannot assume responsibility for submitted work, declare competing interests, or enter into copyright and licensing agreements.

Authors who use generative AI or AI-assisted technologies must disclose the name and version of the tool, the purpose for which it was used, and how its output was reviewed and verified. Authors remain fully responsible for the accuracy, originality, citations, legality, ethical compliance, and integrity of all submitted content.

Principles for the Use of Generative AI

a) Scholarly writing and interpretation: Generative AI must not replace the authors’ critical thinking, scholarly judgement, analysis, interpretation, or original contribution. AI-generated material must not be presented as the authors’ own work without substantial human review, verification, revision, and transparent disclosure.

b) Research data and findings: Generative AI must not be used to fabricate, falsify, or misrepresent data, findings, observations, citations, or conclusions. Where AI tools form part of the research methodology, their use must be described transparently and in sufficient detail for editorial evaluation and, where possible, reproducibility.

c) Statistical and computational analysis: AI tools may assist with statistical or computational analysis where methodologically appropriate. Authors must independently verify all calculations, outputs, assumptions, code, interpretations, and reported statistics.

d) Language and readability: AI-assisted tools may be used to improve language, grammar, clarity, or readability, provided they do not replace scholarly judgement or introduce unverified claims, data, citations, or interpretations. Substantive use must be disclosed, and authors remain responsible for the final manuscript.

e) Images and visual materials: The journal does not accept images, figures, graphical abstracts, or other visual content generated or materially altered using generative AI. AI must not be used to fabricate, manipulate, obscure, or misrepresent primary research data or visual evidence. Where AI-based image generation or processing is itself part of the research methodology, it may be considered only when scientifically justified, transparently described, appropriately validated, and compliant with ethical, legal, copyright, and data-integrity requirements.

f) Confidentiality and data protection: Authors must not upload confidential manuscripts, personal data, sensitive research data, copyrighted materials, or information belonging to third parties into AI systems unless they have the necessary authority and the use complies with applicable ethical, legal, institutional, and data-protection requirements.

g) Editorial assessment: An AI-detection indicator will not, by itself, be treated as conclusive evidence of undisclosed AI use or misconduct. Where concerns arise, the journal will assess the available evidence, inform the authors where appropriate, and provide a reasonable opportunity to respond before reaching an editorial decision.