Generative AI Policy

Purpose and Scope

The SDEWES Journals Generative AI Policy sets out the principles and procedures for the responsible, transparent and accountable use of generative artificial intelligence and AI-assisted technologies in journals published by SDEWES Centre.

This policy applies to all submitted, accepted and published content, including regular articles, review articles, special issue papers, article collections, conference-linked submissions and supplementary materials.

It applies to authors, reviewers, Editors-in-Chief, Associate Editors, Guest Editors, Editorial Board members, SDEWES Centre staff or representatives and other individuals involved in manuscript preparation, research, peer review, editorial assessment, production, publication or post-publication processes.

The policy covers the use of generative AI and AI-assisted technologies in writing, editing, translation, literature organisation, data or image processing, code development, modelling, analysis, manuscript preparation, peer review, editorial work, publication workflows and post-publication assessment.

Related guidance, statement templates, declaration forms and checklists relevant to this policy are listed in Appendix A: Related COPE and International Guidance and Appendix B: Practical Forms, Statements and Checklists.

Definition and General Principles

For the purposes of this policy, generative AI means a system or service that generates or substantially transforms text, images, audio, video, code, data, figures or other content in response to prompts, instructions, examples or other inputs.

AI-assisted technologies means tools that incorporate generative or content-transforming functions into writing, translation, analysis, coding, visualisation, editorial or publication workflows.

Conventional statistical, numerical, modelling, simulation, optimisation, reference-management, formatting or programming tools are not treated as generative AI solely because they contain automated or AI-based functions. However, their use must be reported where they materially generate or transform scholarly content or form a substantive part of the research methodology.

For reviewer, editor or publisher use, an approved tool means a tool expressly authorised by SDEWES Centre or the relevant journal for the intended purpose and used under arrangements that protect confidential information and personal data.

Generative AI may support research and scholarly communication, but must not replace human responsibility, scientific judgement, critical thinking, originality, accountability, peer review or editorial decision-making.

Its use must be guided by transparency, accuracy, integrity, confidentiality, privacy, copyright compliance, human oversight and protection of the scholarly record.

Human Responsibility and Accountability

Authors remain fully responsible for the accuracy, originality, validity, integrity and reliability of all content submitted to SDEWES journals, including content generated, edited, translated, analysed or otherwise supported by generative AI or AI-assisted technologies.

Authors must critically review, verify and, where necessary, correct AI-generated or AI-assisted output before it is incorporated into a manuscript, figure, table, code, dataset, supplementary material, revision or response to reviewers.

Authors must be able to explain and substantiate the content, methods, reasoning, references, data, code and conclusions for which they claim responsibility. Reliance on an AI tool does not excuse an author from providing underlying evidence or responding to questions about the work.

Authors must ensure that AI-assisted content does not contain fabricated references, inaccurate claims, misleading interpretations, plagiarism, inappropriate image or data alteration, copyright infringement, confidential information, unlawfully processed personal data, biased content or unsupported conclusions.

Responsibility cannot be transferred to an AI tool, its provider or another third party. Concerns involving fabrication, falsification, plagiarism or other possible misconduct are assessed under Publishing Ethics and the relevant specialised SDEWES Journals Policies.

Generative AI and Authorship

Generative AI tools and AI-assisted technologies cannot be listed as authors or co-authors because they cannot approve the manuscript, accept responsibility, respond to questions about the work or hold legal rights and obligations.

AI tools must not be assigned authorship, human contributor status or a CRediT contributor role. Human authors remain responsible for content created or supported using such tools.

Where relevant, an AI tool may be identified as software, a research method or a manuscript-preparation tool in the disclosure statement, Methods section or reference list, as appropriate. This does not confer authorship or contributorship.

Authorship and contributorship requirements are governed by Authorship and Contributorship, while copyright and licensing requirements are governed by Copyright & Licensing.

Author Use of Generative AI in Manuscript Preparation

Authors may use generative AI or AI-assisted technologies in a supportive role where the use complies with this policy and the relevant journal’s Guide for Authors.

Permitted supportive uses may include:

improving the clarity, readability or structure of author-written text;
translating author-written material;
organising author-prepared content;
assisting with literature organisation or search-term development;
supporting code development, checking or documentation;
assisting with technical formatting or compliance checks; and
preparing explanatory or illustrative material where permitted.

Authors must independently locate, read and verify all scholarly sources. AI-generated summaries or references must not substitute for the authors’ own assessment of the cited literature.

AI output should not be cited as authoritative evidence for a factual or scholarly claim unless the AI system or its output is itself the subject of the research and the use is reported transparently.

Generative AI must not replace the authors’ scientific judgement, analysis, interpretation, conclusions or original intellectual contribution. Translation or language assistance must not alter the intended scientific meaning.

Disclosure of Generative AI Use by Authors

Authors must disclose substantive use of generative AI or AI-assisted technologies in manuscript preparation, including use to generate or substantially transform text, translations, analyses, interpretations, summaries, figures, tables, code, supplementary material or responses to reviewers.

The disclosure should identify, as applicable:

the tool, model or service used;
the developer or provider;
the version and date or period of use, where available;
the purpose and extent of use;
the manuscript section, research activity or content affected; and
the human review, verification and validation performed.

Use in manuscript preparation must be disclosed in a dedicated manuscript statement in the location required by the journal. Use as part of the research methodology must also be reported in the Methods section or another appropriate section in sufficient detail to support evaluation and reproducibility.

Where generative AI was used both as a research method and in manuscript preparation, both forms of reporting may be required.

Authors must update the declaration where substantive use occurs or changes during revision, production or preparation of responses to reviewers.

A declaration that generative AI was not used may be collected through the submission system or author declaration form, where required by the journal, but does not necessarily need to be published with the article.

Where the submission system does not collect the required information, authors may be required to submit the Generative AI Use Declaration Form listed in Appendix B. Standard manuscript wording is provided in the Generative AI and AI-Assisted Technologies Disclosure Statement Template.

Uses That Do Not Normally Require Disclosure

Tools used solely for basic spelling, grammar, punctuation, formatting, reference-style checking or accessibility support do not normally require disclosure where they do not generate new scholarly content, materially alter scientific meaning, perform analysis or interpretation, or substantially restructure the manuscript.

Standard word-processing, reference-management and technical formatting functions do not normally require disclosure.

Translation, rewriting, summarisation or language-editing functions that materially transform the text or meaning should be disclosed under Section 6.

Where there is reasonable doubt about whether a use is substantive, disclosure is preferred.

Generative AI in Research Methods, Data, Code and Images

Where generative AI or AI-assisted technologies form part of the research methodology, data processing, modelling, simulation, code development, statistical analysis, image processing or figure generation, the use must be described transparently and in sufficient detail for evaluation and, where reasonably possible, reproducibility.

The methodological description should include, where relevant:

tool or model name, version, developer or provider;
date or period of access;
purpose and role in the research;
relevant prompts, parameters or configuration settings;
input data and preprocessing;
generated or transformed outputs;
validation and quality-control procedures;
human oversight; and
known limitations or uncertainty.

AI-generated or synthetic data must be clearly identified and distinguished from observed, measured, experimental or participant-derived data.

Generative AI must not be used to fabricate, invent or misrepresent research data or results. AI-assisted processing must not obscure the relationship between the original data and the reported findings.

Generative AI must not create or alter images presented as primary observed or experimental evidence unless the generation or transformation is itself part of the declared research method. In such cases, the original material, transformation process and validation must be documented and retained where appropriate.

AI-generated illustrative images, graphical abstracts or cover material must be clearly disclosed, comply with journal instructions and satisfy applicable copyright, licensing and permission requirements.

AI-assisted code must be reviewed, tested and documented by the authors. Data, code, model and reproducibility requirements are governed by Research Data and Transparency.

Use involving personal, confidential or human-participant data must also comply with Research Ethics and Privacy. Free or discounted model access, computing credits, vendor support or other relevant in-kind assistance must be disclosed under Competing Interests and Funding Disclosure.

Prohibited or Restricted Uses by Authors

Authors must not use generative AI or AI-assisted technologies to:

fabricate, falsify or misrepresent data, results, images, references, quotations or research records;
create false authorship information, ethics approvals, consent records, reviewer identities, competing-interest declarations or other publication documentation;
conceal plagiarism, duplicate publication, inappropriate image alteration or another integrity concern;
evade peer review, reporting standards, ethical review, data-transparency requirements or journal screening procedures;
impersonate an author, reviewer, editor, participant or other person;
include AI-generated references without independently verifying the existence, accuracy and relevance of each source; or
present AI-generated material as an original human contribution where doing so would be misleading.

Authors must not enter confidential peer-review reports, editorial correspondence, unpublished third-party material, personal data, sensitive data or copyrighted material into a generative AI system unless they have the necessary rights, permissions, safeguards and lawful basis for the intended processing.

Authors must consider the tool’s data-retention, reuse, training, sharing and security practices before entering any non-public material.

Generative AI Use by Reviewers

Manuscripts, supplementary materials, review reports, author responses, reviewer identities and editorial correspondence are confidential.

Reviewers must not upload confidential peer-review material to an external generative AI tool unless the tool has been expressly approved by SDEWES Centre or the relevant journal for that purpose.

Peer review must not be delegated to generative AI. Reviewers remain personally responsible for the analysis, reasoning, accuracy, fairness, independence and integrity of their reports and recommendations.

Any permitted AI use must be limited to a supportive function and must not replace the reviewer’s own expert assessment. Permitted substantive use must be disclosed to the editor, except for basic spelling, grammar or formatting functions covered by Section 7.

Reviewers must not upload manuscripts to AI-detection services or other external screening tools unless explicitly authorised by the journal. Suspected inappropriate AI use should instead be reported confidentially to the editor.

Detailed confidentiality, reviewer-conduct and peer-review requirements are governed by Peer Review Policy and Privacy.

Generative AI Use by Editors and SDEWES Centre

Editors, Guest Editors, Editorial Board members and SDEWES Centre staff or representatives must protect manuscripts, reviewer reports, author responses, reviewer identities, editorial deliberations, ethics-case information and unpublished materials.

Confidential content must not be uploaded to an external generative AI system unless the tool has been expressly approved for the intended purpose and the applicable confidentiality, privacy, copyright and data-protection safeguards are in place.

Generative AI must not replace independent editorial judgement, reviewer assessment, ethics evaluation or final decision-making. A human editor authorised by the journal remains responsible for every editorial decision.

Approved AI-assisted tools may be used for limited administrative, technical, accessibility, metadata, production or language-support purposes where human oversight is maintained and confidential information is protected.

AI-generated assessments, summaries, reviewer recommendations, integrity scores or screening results must not serve as the sole basis for:

reviewer selection;
acceptance or rejection;
findings of misconduct;
corrections, expressions of concern or retractions; or
decisions on appeals or complaints.

Editorial independence is governed by Editorial Independence, peer-review procedures by Peer Review Policy, and confidential information by Privacy.

The current policy’s reference to AI-supported identification of reviewer expertise should be removed or covered by the broader rule above. Reviewer selection must remain a human editorial responsibility.

Confidentiality, Privacy, Copyright and Data Protection

Any use of generative AI or AI-assisted technologies must protect confidentiality, privacy, personal data, intellectual property, copyright and other legal or ethical obligations.

Before using a tool, the user must consider whether the system stores, reuses, shares, transfers, trains on or otherwise processes submitted content and whether the intended use is compatible with the person’s rights and obligations.

A paid, private or enterprise account is not automatically considered approved for confidential reviewer, editor or publisher use. Approval must apply to the specific tool and intended purpose.

Confidential manuscripts, unpublished data, reviewer identities, editorial correspondence, personal or sensitive data and third-party material must not be entered into a system where the applicable safeguards, permissions or rights are absent.

Privacy and personal-data requirements are governed by Privacy, copyright and licensing by Copyright & Licensing, and peer-review confidentiality by Peer Review Policy.

Undisclosed, Inappropriate or Suspected Use

A concern about possible AI use must be assessed fairly, confidentially and proportionately. The journal should assess the nature and effect of the use rather than treating any use of generative AI as misconduct in itself.

Writing style, unusual phrasing, reviewer suspicion or an AI-detection result must not serve as the sole evidence that generative AI was used or that misconduct occurred. Confidential manuscripts must not be uploaded to external AI-detection systems unless the tool is approved for that purpose.

The journal may request information reasonably necessary for the assessment, including clarification of:

the tool or model used;
the purpose and extent of use;
affected manuscript sections or research activities;
relevant version or date information;
author review and validation;
underlying data or code; and
relevant prompts or outputs where necessary and proportionate.

Editors should not request or retain prompts, outputs or other records that contain unnecessary personal, confidential, proprietary or third-party information.

The assessment should distinguish between:

permitted use that did not require disclosure;
permitted use that should have been disclosed;
incomplete or unclear disclosure;
inappropriate but remediable use; and
use involving fabrication, falsification, plagiarism, confidentiality breaches, false declarations or another serious integrity concern.

Before publication, the journal may request clarification, revised text, corrected references, updated declarations, additional validation or removal of inappropriate content. Rejection or discontinuation of handling may be appropriate where the integrity concern cannot be remedied or where serious misconduct is suspected.

After publication, any amendment to the scholarly record must be considered under Corrections, Retractions and Withdrawals. Serious integrity concerns are assessed under Publishing Ethics and one or more relevant specialised SDEWES Journals Policies.

Personal data and confidential information must be handled under Privacy. Formal appeals or complaints concerning the journal’s handling of the matter are assessed under Editorial Decision Appeals and Complaints.

Policy Review and Updates

SDEWES Centre may review and update this policy periodically to reflect developments in generative AI technologies, scholarly publishing, publication ethics, research integrity, peer review, copyright, privacy, data protection, open science, legal requirements and international standards.

Updated policies apply from the date of publication on the SDEWES Journals Policies website unless otherwise stated.

Appendix A. Related COPE and International Guidance

COPE Council. COPE position - Authorship and AI tools

Relevant to the rule that AI tools cannot be listed as authors because they cannot take responsibility for submitted work, manage competing interests or manage copyright and licensing.

COPE Topic Discussion. When editors suspect AI

Relevant where editors need to assess suspected AI use, possible misconduct, disclosure concerns or integrity risks.

COPE case: Reviewer suspects use of AI by an author

Relevant where editors or reviewers suspect undisclosed or inappropriate AI use in a submitted manuscript.

COPE Council. Editors suspect that reviewers are using artificial intelligence

Relevant where review reports may have been substantially generated by AI, expert review may have been delegated, or confidentiality and reviewer responsibility may have been compromised.

COPE Council. Emerging AI dilemmas in scholarly publishing

Provides broader discussion of accountability, detection limitations, confidentiality and evolving AI-related risks in scholarly publishing.

Appendix B. Practical Forms, Statements and Checklists

The documents below provide practical support for author disclosure, preparation of manuscript statements and proportionate assessment of potentially undisclosed, inappropriate or misleading generative AI use. Peer-review confidentiality checks are listed in Appendix B of Peer Review Policy.

Generative AI Use Declaration Form

Required author submission form, where required – Used by authors at submission to declare whether generative AI or AI-assisted technologies were used and, where applicable, to describe the tool, purpose, manuscript section affected and author verification. This form may be required where the submission system does not collect these declarations directly.

Generative AI and AI-Assisted Technologies Disclosure Statement Template

Author guidance template – Used by authors to prepare the required manuscript statement where generative AI or AI-assisted technologies were used in preparing, developing, translating, analysing or substantially editing the manuscript.

Generative AI Concern Checklist for Editors

Internal editor checklist – Used by the journal’s Editor-in-Chief and SDEWES Centre when undisclosed, inappropriate, misleading or potentially problematic generative AI use is suspected before or after publication.