Artificial intelligence cannot be an author
AI tools cannot be listed as authors. They cannot consent to submission, approve the final version, or take responsibility for the work. Authors remain responsible for the accuracy, originality, references, and integrity of every part of the manuscript.
Obligations of authors: disclosure
Substantive use of generative AI in writing, analysis, coding, or image preparation must be disclosed, identifying the tool and its purpose. Authors must verify all outputs and must not submit fabricated data, invented citations, or misleading images.
Routine spelling and grammar correction need not be disclosed, provided it does not change scholarly content, data, or argument. Where there is doubt, disclose. A reference produced by a tool and not checked against the primary source is treated as a fabricated citation.
Obligations of reviewers: confidentiality
Reviewers must not upload confidential manuscripts, personal data, or unpublished results to external AI services without appropriate authorization and confidentiality safeguards. A manuscript under review is a confidential document, and transferring it to a third-party service breaches that confidence. A reviewer unable to complete an assessment without such assistance should decline the invitation instead. AI output does not replace expert peer review.
Obligations of editors: limits on use
Editors are bound by the same confidentiality and must not upload manuscripts, review reports, or author correspondence to external AI services. An editorial decision is made by a named human editor and is never delegated to an automated tool; AI output does not replace editorial judgement.
Tools used by the journal
Similarity screening during editorial assessment may be supported by ORCA, and copyediting and proofreading may be supported by EIGO, as set out in the Submission Process. In both cases the output is decision support only: all editorial judgments remain the responsibility of SJS editors, and an edited version is returned to the corresponding author for approval before the galley proof.
Breaches
Undisclosed substantive use, fabricated data or references, or presenting generated material as authentic is handled under Publication Ethics and Authorship, and may lead to correction, retraction, or notification of the authors’ institution.
Read together with Publication Ethics and Authorship, Similarity and Plagiarism Policy and Peer Review Policy. This policy is reviewed at least once a year.