Responsible AI Policy

1.Where AI is used

• AI Integrity Check: text similarity, image duplication and alteration, reference lookup and AI-text indicators.
• Reviewer Match: suggesting reviewers from topic and publication history, and screening possible conflicts of interest.
• Submission Desk: parsing references and reading metadata from uploaded files.

2.Editors decide

AI checks flag possible problems. Editors investigate and decide. No acceptance, rejection, reviewer invitation or misconduct finding is ever made automatically.

3.A flag is not proof

A flag is never treated as proof of misconduct. Each flag in a report shows its evidence and its limits, so that the editor can judge context. Similarity with an author's own cited methods, for example, is reported differently from similarity in results.

Before acting on a flag, journals are expected to give authors the chance to explain, in line with their own policies and guidance from bodies such as COPE.

4.AI-text indicators

AI-text indicators are shown as uncertain signals only. Detection of generated text is unreliable, particularly for authors writing in a second language and for heavily edited text. These indicators are never the sole basis for rejection, and the report states this beside every result.

5.Confidentiality of manuscripts

Unpublished manuscripts are confidential. They are never used to train AI models and never shown to anyone outside the editorial process. Manuscript content processed by a check is used only to produce that manuscript's report.

6.Fairness in reviewer suggestions

Reviewer Match ranks candidates by fit to the manuscript, not by name, gender or country. Editors can filter for region and career stage to broaden their reviewer pool, and always choose who to invite.

7.Reviewer identities

Reviewer identities are protected according to the journal's review model. AI tools do not reveal reviewer names to authors under single-blind or double-blind review.

8.Supporting, not replacing, journal policy

The platform supports, but does not replace, each journal's own ethics policies. Journals set their own rules on authors' use of AI tools, and Submission Desk can ask authors to declare such use.

9.Questions and concerns

If you believe a check has produced a misleading result, or you have a concern about how AI is used, write to us through the contact page. We review every report and use them to improve the checks.