Policy on the Use of Artificial Intelligence

Artificial intelligence (AI) tools, including generative language models and related technologies, may support certain stages of research and scholarly communication. Their use, however, must remain consistent with the fundamental principles of research integrity: transparency, human accountability, scholarly responsibility, and traceability of contributions.

Our approach is grounded in the principle that AI may augment — but must not replace — human intellectual judgement. When used responsibly and appropriately, AI technologies may facilitate scientific discovery, improve access to scholarly communication, and support the quality and efficiency of editorial processes. Nevertheless, all outputs generated or supported by AI require critical human evaluation, factual verification, and editorial scrutiny. Meaningful human oversight must therefore remain integral at every stage of research production, manuscript preparation, peer review and editorial assessment.

Authors, editors, and peer reviewers retain full and individual responsibility for the content they produce and for all scholarly, editorial, and evaluative decisions made throughout the publication process, irrespective of whether AI tools were employed.

This policy is informed by the recommendations of the Committee on Publication Ethics (COPE) and is intended to evolve alongside developments in academic practice, technological innovation, and applicable regulatory frameworks.

Authors

Permitted Use

Authors may use AI tools to:

• improve the linguistic quality of their text (grammar correction, stylistic enhancement, translation);
• assist writing (structuring the manuscript, reformulating passages);
• support certain stages of analysis, provided that the use of AI is described in detail in the methodology section of the article with sufficient precision to ensure reproducibility.

Given that translation is a particularly important issue for a language sciences journal, any full or substantial translation carried out using an AI tool must be declared according to the procedures set out in the Transparency Requirements section below, and the translated text must undergo thorough human revision.

Transparency Requirements

Any use of AI that materially contributes to the generation, drafting, revision, translation, analysis, visual production, or editing of manuscript content must be explicitly declared by the authors at the time of submission and, where appropriate, within the published article.

For guidance, the following uses require disclosure:

• reformulating or drafting entire passages;
• full or partial translation of an article;
• generation of figures or diagrams;
• assistance with data analysis or coding;
• generation or reorganisation of bibliographic references.

By contrast, the following uses do not require disclosure:

• use of a standard spell checker;
• occasional grammar checking;
• use of autocomplete functions integrated into word-processing software.

The declaration must appear in a dedicated section of the article (before or after acknowledgements, or before the references) and include:

• the name of the tool used (and version, if known);
• the type of use (writing, translation, analysis, image generation, etc.);
• the extent of human involvement in verifying and revising the generated content.

Example declaration:

"The authors used [tool name, version] to [precise description of use, e.g., improve the stylistic fluency of the French text / translate an initial version of the manuscript from English into French / assist thematic coding of interview data]. All content generated in this way was reviewed, corrected, and validated by the authors, who assume full responsibility for it."

Responsibility

Authors bear full scientific, ethical, and legal responsibility for all submitted content, including portions produced with AI assistance. They remain solely accountable for the originality, accuracy, integrity, and attribution of their work.

Accordingly, authors must:

• rigorously verify the scientific accuracy of any AI-generated content, ensuring in particular the absence of factual errors, fabrications, biases, or nonexistent references;
• guarantee the originality of the text and the integrity of the data presented.

Because an AI system cannot be held accountable, no error or inaccuracy may be attributed to the tool used.

Restrictions

AI systems may under no circumstances be listed as authors or co-authors of an article submitted to the journal. AI tools cannot qualify as authors and must not be credited with authorship.

The use of AI to generate, modify, or fabricate scientific data without transparent disclosure is prohibited.

With regard to images, three situations must be distinguished:

• AI-generated purely illustrative or conceptual images (diagrams, synthetic figures) must be clearly identified as such in their captions.
• Data visualisations (graphs, charts) produced with AI assistance must indicate the tool used.
• The use of AI to create, modify, or alter images that claim to represent empirical data (source reproductions, screenshots, field photographs, etc.) is strictly prohibited and constitutes a breach of research integrity.

Intellectual Property

Authors guarantee that they hold the rights to all submitted content, including portions produced with AI assistance.

The journal cannot be held liable for copyright issues arising from the use of AI-generated content and reserves the right to withdraw a published article if an intellectual property violation is established.

Reviewers

Confidentiality

Manuscripts submitted for review are confidential documents. To preserve confidentiality, reviewers must not upload, copy, or enter any part of a manuscript — including short excerpts — into an external AI tool or generative platform.

Limited Use

Reviewers may use AI tools only in a very limited manner, in particular for the purpose of improving the linguistic quality of their own review reports.

Responsibility

Review reports must reflect the reviewer's personal scientific judgement and be based on independent critical evaluation. AI may not under any circumstances be used to formulate an assessment, recommendation, or decision regarding the manuscript under review.

Reporting Suspected Undisclosed AI Use

If a reviewer suspects that a manuscript has been largely written using AI without disclosure, they are encouraged to report this to the editor in their review or via separate communication. The editor will determine the appropriate course of action, which may include requesting clarification from the authors.

Editors

Role of AI

AI tools may be used by the editorial team for administrative or technical purposes, such as:

• sorting submissions;
• formally checking manuscript compliance;
• detecting textual similarities.

Such uses are considered non-decision-making editorial assistance. AI may support administrative or linguistic aspects of editorial work, but must not substitute for expert evaluation.

Detection of AI-Generated Content

The journal reserves the right to use AI-content detection tools during the submission evaluation process. Where applicable, the results of such checks constitute only one element of assessment and cannot alone justify rejection. It shall fall to the editor to evaluate the concern and respond accordingly, including, where warranted, requesting that the authors provide clarification.

Limitations

Editors do not use AI to make editorial decisions, evaluate the scientific quality of manuscripts, or replace human judgment at any stage of the decision-making process.

Editorial Responsibility

Editorial decisions must be based on human expertise and peer-review reports. The Editor-in-Chief remains responsible for ensuring implementation of this policy.

Compliance and Sanctions

Failure to comply with this policy may result, depending on the seriousness of the situation, in:

• a request for clarification or correction addressed to the authors;
• rejection of the manuscript;
• retraction of an already published article, accompanied by a retraction notice in accordance with COPE guidelines.