ST ANTONY'S INTERNATIONAL REVIEW
Policy on the Use of Generative Artificial Intelligence 2026-2027
1. Purpose and scope
This policy sets out how authors, peer reviewers and editors of St Antony's International Review (STAIR) may and may not use generative artificial intelligence (AI) tools, such as large language models and AI image generators, in the preparation, review and editing of material submitted to the journal. It aims to protect the confidentiality of the peer review process, the integrity of published scholarship, and the accountability of everyone involved in it.
The policy applies to all submissions, including research articles, review essays and book reviews, and to all stages of the editorial process.
2. Authors
2.1 Permitted uses
STAIR supports the responsible use of generative AI tools by authors, provided that use is transparent and subject to human oversight. Permitted uses include:
• idea generation and exploration;
• copyediting, language improvement and translation;
• classifying or organising literature;
• coding assistance; and
• producing data visualisations or conceptual illustrations, such as flow diagrams.
Where any of these uses involves licensed or copyrighted material, the rules in section 5 also apply.
2.2 Prohibited uses
Authors must not use generative AI tools to replace their core responsibilities as authors. In particular, the following are not permitted:
• generating text or code that is included in the submission without rigorous revision by the author;
• generating synthetic data to substitute for missing data without a robust and disclosed methodology;
• generating content that is inaccurate, including abstracts or supplementary materials;
• creating or manipulating research outputs, including images, maps or data, by augmenting, concealing, moving, removing or introducing any feature; and
• producing fabricated, misattributed or unverifiable citations.
2.3 Disclosure
Any use of generative AI tools beyond basic spelling and grammar checking must be acknowledged in a disclosure statement placed before the references. The statement must give:
• the full name and version of the tool used;
• how it was used; and
• the reason for its use.
Example: "The author used ChatGPT (GPT-4o) to improve the readability and language of Sections 2 and 3. The author reviewed and edited all output and takes full responsibility for the content of this article."
2.4 Authorship and accountability
Generative AI tools cannot be listed as an author or co-author, because they cannot take responsibility for the submitted work or enter into copyright and licensing agreements. Authors remain fully accountable for the originality, accuracy and integrity of their submission, including any content produced with AI assistance. Any use of AI must comply with the journal's policy on plagiarism.
3. Peer reviewers
Manuscripts under review are confidential. Uploading any part of an unpublished manuscript to a generative AI tool may breach that confidentiality and the author's rights over their work. Accordingly, peer reviewers:
• must not upload unpublished manuscripts, or any files, images or information from them, into generative AI tools;
• must not use generative AI to analyse or summarise a submission, or any part of it;
• must not use generative AI to write or generate their review report; and
• may use generative AI only to improve the language of a review they have written themselves, provided no confidential manuscript content is entered into the tool.
Reviewers remain responsible at all times for the accuracy and integrity of their reviews.
4. Editors
Editors must not upload unpublished manuscripts, or any associated files, images or information, into generative AI tools. Editors who wish to use generative AI for any other editorial task, such as drafting correspondence, should first consult the Editor-in-Chief. Editorial decisions must be made by human editors.
5. Licensed and copyrighted content
Much of the material that authors, reviewers and editors work with, such as journal articles, e-books and databases, is licensed by libraries for limited purposes. Anyone using such material with AI tools in connection with STAIR must follow these rules:
• Purpose. Licensed content may be used with AI tools only for non-commercial research, teaching and learning. Commercial use of AI-generated content, or of platforms that use licensed content, is prohibited.
• Open AI systems. Licensed content must not be used with publicly accessible AI systems where it would be reused or shared.
• Third-party AI tools. Licensed content may be used with third-party AI tools only if it is not used to train the system and is not shared with third parties. Users should check a tool's settings and terms, for example by switching off training on their inputs, before uploading anything.
• Closed institutional systems. Licensed content may be processed and analysed in an AI system that is operated locally or in a controlled environment and is accessible only to the licensing institutions. It must not be shared with third parties or used to train publicly accessible systems.
• Publishing AI-generated results. AI-generated results may be included in work submitted to STAIR, except where they contain licensed content or derive from CC BY-NC-ND or other No Derivatives sources.
6. Breaches of this policy
Suspected breaches will be investigated by the editorial board. Depending on the circumstances, outcomes may include a request for revision or disclosure, rejection of a submission, removal of a reviewer from the journal's reviewer pool, or correction or retraction of a published article.
7. Review of this policy
Generative AI tools and publishing norms are developing quickly. The editorial board will review this policy at least once a year.
