1. Purpose and scope
This policy governs the use of artificial intelligence (AI) tools by authors, reviewers, editors, editorial and advisory board members, and guest editors at every stage of submission, evaluation, revision, and publication in Lectio Socialis.
It is aligned with the COPE position statement Authorship and AI tools (2023), the ICMJE Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals (updated January 2026), in particular Section V, “Use of Artificial Intelligence in Publishing”, the WAME Recommendations on Chatbots and Generative Artificial Intelligence in Relation to Scholarly Publications (revised 31 May 2023), the research-integrity guidance of Cambridge University Press, Elsevier, Springer Nature, Taylor & Francis and Sage, and the COPE/DOAJ/OASPA/WAME Principles of Transparency and Best Practice in Scholarly Publishing (version 4, 2022). It is applied consistently with the journal’s obligations under the TÜBİTAK ULAKBİM TR Dizin journal evaluation criteria, which do not themselves address artificial intelligence.
Where an author's institution, funder, or national regulation imposes stricter requirements, the stricter requirement applies.
1.4 Commencement. This policy applies to manuscripts submitted on or after 1 October 2026. For manuscripts already under consideration on that date, authors are asked to supply the declaration at the next revision stage; the enforcement provisions of Section 13 are not applied retrospectively.
2. Definitions
2.1 Generative AI (GenAI). Any system that produces text, translation, code, images, figures, data, or references in response to user prompts — including large language model chatbots, AI writing and rewriting assistants, AI image generators, and AI-based literature or research assistants.
2.2 Tools outside the disclosure requirement. Spelling, grammar and punctuation checkers, reference-management software, similarity-detection software, standard statistical and qualitative-analysis packages used in their conventional functions, non-generative transcription tools, and AI-assisted search within bibliographic databases as defined in Section 6.
2.3 The functional test. What matters is the function performed, not the product name. Where a tool moves beyond correcting an author's own sentences into rewriting, restructuring, translating, summarising, or producing new content, it falls under Section 2.1 and must be declared.
3. Core principles
3.1 Human accountability. AI is a tool. Full scholarly, ethical, and legal responsibility for every part of a submission rests with the human authors. This includes responsibility for originality: generative models can reproduce protected text from their training data, and authors warrant that AI-assisted output does not reproduce third-party copyrighted material without permission and appropriate attribution.
3.2 Transparency. Any use falling under Section 2.1 is declared openly and specifically.
3.3 Verifiability. Every claim, figure, quotation, and reference produced with AI assistance must be verified by the authors against the original source.
3.4 Confidentiality. No confidential information is shared with any AI system at any stage. This covers manuscripts under consideration, reviewer reports and identities, editorial correspondence, unpublished data, and identifiable personal data.
3.5 Non-discrimination. AI-detection software is unreliable and produces false positives for authors writing in English as an additional language. No editorial decision is taken on the basis of detection output alone.
3.6 Proportionality. Declared, legitimate, well-supervised use of AI is not in itself a ground for rejection; undeclared or unverified use is.
4. Authorship
4.1 AI tools cannot be listed as an author or co-author and cannot be cited as an author, because they cannot take responsibility for the work, approve the final version, declare conflicts of interest, or respond to post-publication queries.
4.2 All listed authors must satisfy the journal's human authorship criteria and are jointly accountable for AI-assisted portions of the text as if they had written them unaided.
4.3 AI tools may be named in the AI declaration or acknowledgements as instruments, never as contributors, and are not assigned CRediT roles.
5. Uses permitted, subject to declaration
- Language editing, style improvement, and readability of the authors' own draft.
- Translation into English of text originally drafted by the authors.
- Formatting of text, tables, and references.
- Summarising or organising literature that the authors have themselves read and verified (see Section 6).
- Assistance with code, scripts, or statistical syntax, provided the output is checked and the analysis is reproducible.
Structuring, outlining, or brainstorming, provided the substantive argument and interpretation remain the authors' own.
Where structuring or outlining assistance is declared, the declaration states what was structured. The test the journal applies is whether the authors can explain and defend the argument, the selection of evidence, and the interpretation independently of the tool. Where they cannot, the use falls under Section 7 rather than this section.
6. AI-assisted literature search and citation integrity
6.1 Recognition. AI-assisted discovery is now embedded in standard research infrastructure, including tools built into major bibliographic platforms. Their use is legitimate and is not in itself a matter for declaration. What the journal regulates is not the act of searching but the integrity of what is cited.
6.2 The governing distinction. A tool that retrieves real, indexed bibliographic records functions as a finding aid and is treated like any other database search. A tool that generates prose about the literature — summaries, syntheses, "related work" paragraphs, reference lists composed from the model's parameters — produces text, not evidence, and falls under Sections 2.1 and 7.
6.3 The consultation rule. Authors may cite only sources they have themselves consulted in the form in which they cite them. No citation, quotation, page reference, date, or attribution of an argument may rest on an AI-generated summary, snippet, or abstract-level paraphrase. Every reference is verified against the original record before submission.
6.4 Synthesis remains the authors' work. AI-generated summaries may inform the authors' reading, but the selection of the literature, the structure of the review, and the critical assessment of the field must be the authors' own intellectual work.
6.5 Specifically not permitted.
- Reference lists produced by a model answering from its parameters rather than from retrieved records (Section 6.2), or any reference not verified against an indexed record or the source itself.
- Citing works retrieved but not read, or padding a bibliography with AI-suggested items.
- Citing retracted, withdrawn, or fabricated items surfaced by a tool without verification.
- Basing a quotation, a translated passage, or an attributed claim on a tool's paraphrase.
- Delegating inclusion or exclusion decisions in a review to an AI tool without documented human screening.
6.6 Review articles and systematic evidence synthesis. Where AI tools contribute to searching, screening, deduplication, or data extraction, the Methods section must report the tool and version, the dates of use, the queries or prompts, the known coverage limits of the tool, and the human verification procedure. Because the coverage and ranking logic of AI search tools is generally proprietary and not reproducible, at least one conventional, fully reported database search strategy must also be provided.
7. Uses not permitted
- Generating or fabricating data, results, cases, respondents, interview material, or archival sources.
- Producing references, quotations, page numbers, legal citations, or statistics that have not been verified against the original source (see Section 6).
- Generating, altering, or enhancing figures, photographs, maps, or other visual material, except where such material is itself the object of study (Section 9). Ordinary chart production from the authors' own verified data is permitted and should be described.
- Presenting AI-generated analysis, interpretation, theoretical argument, or literature review as the authors' original intellectual contribution.
- Entering confidential or personal third-party material — unpublished manuscripts, interview transcripts, fieldwork notes, identifiable personal data — into AI tools without a lawful basis under applicable data protection legislation (including the EU General Data Protection Regulation and, in Türkiye, Law No. 6698 on the Protection of Personal Data) and the consent of the data subjects.
- Embedding hidden instructions or prompts in the manuscript file, for example concealed text addressed to AI systems used in screening or review. This is treated as research misconduct and results in immediate rejection.
- Using paraphrasing or rewriting tools to conceal plagiarism or to evade similarity detection.
8. Declaration requirements for authors
8.1 Every submission must carry a section headed "Declaration of Generative AI Use" in two places: on the title page, and after the reference list in the main manuscript. The two must be identical in substance. Because the main manuscript is the file sent to reviewers, the version there must not identify the authors, their institution, or any service acting on their behalf; where naming a tool or provider would do so, the title-page version carries that detail and the manuscript version records the function performed and the extent of human verification. The declaration after the reference list is retained in the published version.
At submission, the corresponding author confirms compliance with this policy through the mandatory generative-AI item in the submission checklist. The declaration in the manuscript is the journal’s record of AI use and the version seen by readers; no separate form is required. Authors may use the model statements in 8.2 as they stand, adapting the bracketed details to their own case.
8.2 Where AI has been used, the declaration must state: the tool name and version/model, the provider, the date(s) of use, the purpose, the sections of the manuscript affected, and the extent of human verification.
8.2.1 Model statement (use): "During the preparation of this manuscript, the author(s) used [tool, version, provider] in [month, year] for [purpose, e.g. language editing of the introduction and discussion]. All output was reviewed and edited by the author(s), who take full responsibility for the content of the publication."
8.2.2 Model statement (no use): "The author(s) declare that no generative artificial intelligence tools were used in the preparation of this manuscript."
8.3 Where AI has been used in research design, data collection, corpus construction, coding, or analysis, this must additionally be described in the Methods (or equivalent) section in sufficient detail for replication: model and version, dates, parameters, sampling procedure, and the prompts used. Prompts should be supplied in an appendix or in an open repository wherever feasible.
8.4 Disclosure obligations continue through revision. Authors must update the declaration if AI is used at the revision stage.
8.5 Work carried out on the authors’ behalf. Where a translation service, a language-editing service, or any other third party acting for the authors has used generative AI, the authors remain responsible for the result and declare that use on the same terms as their own.
9. AI systems and AI output as objects of research
Studies whose subject matter is AI output — for example, discourse analysis of model-generated text, or experiments on model behaviour — are expressly welcome and are not restricted by Section 7, provided that the model, version, access date, prompt set, and sampling strategy are fully documented, that the corpus is deposited in an open repository where licensing permits, and that all interpretation is the authors' own.
10. Reviewers
10.1 Manuscripts under review are confidential and unpublished intellectual property. Reviewers must not enter, upload, or paste any part of a manuscript — text, title, abstract, tables, figures, data, or supplementary files — into generative AI tools or any other third-party system.
10.2 Generative AI must not be used at any stage of the review: not to summarise or analyse the manuscript, not to generate or draft a review report, and not to edit a report already written. Review reports must be wholly the reviewer's own work and must express the reviewer's own expert judgement.
10.3 A reviewer who suspects undeclared AI use in a manuscript reports the concern, with reasons, to the handling editor, and does not run detection software independently.
10.4 Acceptance of a review invitation constitutes acceptance of this section, and the review form carries a corresponding confirmation.
10.5 Where a handling editor has reason to believe that a review report was produced with generative AI, the editor may set the report aside and commission a replacement review, recording the reason in the editorial file. The judgement is the editor’s and is not made on the output of detection software. Indicators that support it include a report that cites no specific content of the manuscript; that attributes to the manuscript arguments, sources, sections, tables or page references it does not contain; that is generic enough to fit any manuscript in the field; or that recommends a decision without stating a reason for it. The reviewer is given an opportunity to respond before any record is made against them.
11. Editors and the editorial office
11.1 Editors are bound by Section 3.4 without exception. Manuscripts under consideration, reviewer reports, reviewer identities, and editorial correspondence are never entered into any AI system.
11.2 Editorial decisions, decision letters, and desk-rejection rationales are produced by human editors. AI is not used to make, or to draft the substance of, any editorial decision, and cannot replace editorial judgement.
11.3 After a manuscript has been accepted, editors may use AI tools in the production stage for language editing, copy-editing, consistency checking, formatting, and reference styling, subject to all of the following conditions:
- the manuscript has passed final editorial decision and is no longer in review;
- no reviewer report, reviewer identity, or item of editorial correspondence is entered into the tool;
- no unpublished third-party data, identifiable personal data, or material the authors have asked to remain restricted is entered into the tool;
- the tool's terms do not permit retention of the input or its use for model training;
- every change is checked by a human editor, and no substantive change to argument, data, or meaning is made;
- the corresponding author receives the edited text for approval at proof stage and may reject any change;
- the use is recorded in the editorial file.
11.4 Any other use of AI by an editor is reported to the Editor-in-Chief.
11.5 Editorial board members handling their own submissions remain subject to the journal's conflict-of-interest procedure; this policy does not displace it.
12. The journal's own use of AI and screening tools
12.1 Lectio Socialis may use similarity-detection and research-integrity screening tools on submitted manuscripts. Authors are informed of this at submission.
12.2 The output of AI-detection software is treated as indicative only. It is never the sole basis for rejection or for an allegation of misconduct. Where a concern arises, the corresponding author is given the opportunity to explain before any decision is taken.
12.3 Any use of AI in the journal's production workflow is limited to the conditions in Section 11.3, is recorded in the editorial file, and is disclosed to the corresponding author at proof stage.
13. Breach of policy
The journal applies a graduated response, following the relevant COPE flowcharts:
- Request for clarification from the corresponding author.
- Request for a corrected or completed declaration, or for revision.
- Desk rejection or rejection where the breach is material or the explanation inadequate.
- After publication: correction, expression of concern, or retraction, and where warranted, notification of the authors' institution.
Fabrication of data or sources, concealed plagiarism, and hidden prompt injection are treated as research misconduct rather than as declaration failures.
An author may contest a decision taken under this policy by writing to the Editor-in-Chief within thirty days of being notified. The appeal is considered by an editor who was not involved in the original decision, and the outcome is given in writing.
14. Review of this policy
This policy is reviewed at least annually by the Editorial Board in the light of developments in COPE, ICMJE, DOAJ, and TR Dizin guidance.