Artificial intelligence (AI) policy

AI tools cannot be authors

Consistent with COPE's position statement on AI tools and the ICMJE recommendations, large language models and other AI tools do not qualify for authorship and must not be listed as authors or co-authors, nor cited as authors. Authorship carries responsibilities that only humans can hold: accountability for the accuracy and integrity of the work, the ability to approve the final version, and the capacity to respond to questions and manage conflicts of interest. An AI tool can do none of these.

Private files in the Lumora workbench

When the file workbench is activated, uploads are extracted locally and the private extraction is encrypted. Review the extracted passages before separately consenting to MiniMax processing. Files and reports expire after 24 hours, or after 30 days if optional history is activated and you select it. You can delete your private content earlier. A request already sent to the provider cannot be recalled. Provider retention depends on the verified provider arrangement; Lumora does not promise zero provider retention or immediate erasure from protected backups. These files are not used for training or optional quality evaluation.

Permitted uses by authors — with disclosure

For author preparation, Lumora tools accept only your own non-sensitive drafts and notes that you are authorized to process. Remove personal identifiers and obtain the required consent before external AI processing. Do not upload confidential material belonging to others, copied reviewer reports or editorial correspondence; use your own non-sensitive notes about feedback when preparing a response.

Authors may use AI tools as assistive technology, provided every output is checked and verified by the authors and the use is disclosed. Permitted uses include:

• Language editing and grammar — improving readability, style, spelling, and grammar of text the authors wrote.
• Translation assistance — drafting or refining translations, which the authors then verify for accuracy and nuance.
• Literature discovery — identifying potentially relevant publications, provided authors read and verify every source before citing it.
• Routine assistance — formatting references, summarising the authors' own notes, or drafting code that the authors test and validate.

In all cases, the authors — not the tool — are responsible for the final text, every citation, and every claim.

Prohibited uses by authors

• Generating or altering research data or results. AI must not be used to create, fabricate, impute, or manipulate data, statistical outputs, or findings presented as real.
• Fabricating references. Citing sources suggested by an AI tool without verifying that they exist and support the claim is a form of fabrication. Every reference must be checked against the original publication.
• Generating or manipulating images of results. AI-generated or AI-altered figures, micrographs, gels, blots, medical images, or other visual evidence of results are not acceptable. Legitimate, disclosed AI-assisted schematic illustrations may be considered case by case where they do not represent research data.
• Undisclosed substantive writing. Using AI to generate substantial scholarly content — analysis, interpretation, discussion — without disclosure.

How authors must disclose AI use

Any use of generative AI beyond basic spell-checking and reference-manager software must be disclosed at submission and in the manuscript itself:

• If AI was used in the research process (for example, data analysis assistance or coding), describe it in the Methods section, including the tool name, version, provider, and how outputs were validated.
• If AI was used in writing or editing, state it in the Acknowledgements (or a dedicated declaration section), for example:

Standard grammar checkers and citation managers that do not generate content do not require disclosure.

Rules for reviewers

Submitted manuscripts are confidential documents. Reviewers must not upload manuscripts, or any part of them, to public or third-party AI tools — doing so breaches confidentiality, may violate the authors' rights in unpublished work, and can expose sensitive data. In addition:

• Peer review is a personal, expert judgement. Reviewers must not use AI tools to generate their assessment or recommendation.
• Limited help with language and organization is allowed only for the reviewer’s own non-confidential comments, with no manuscript excerpts. This includes the Lumora reviewer coach. The tool must not add scientific criticisms or generate an assessment or recommendation; those remain entirely the reviewer’s own.
• Reviewers who suspect undisclosed AI use in a manuscript should raise it confidentially with the handling editor rather than acting on it themselves.

See our Peer Review Policy for reviewers' broader confidentiality obligations.

Responsible use by editors

• Editors must not upload submitted manuscripts, reviewer reports or editorial correspondence to public or third-party generative AI tools, including the Lumora preparation workbench. Acting as a reviewer or editor does not confer permission to share the authors’ confidential work.
• Editorial decisions are human decisions. AI must not be used to determine acceptance or rejection.
• Editors may use vetted, privacy-respecting tools for administrative screening (for example, similarity or image-integrity checks) as an aid to — never a substitute for — editorial judgement.
• Editors handle suspected undisclosed AI use following COPE guidance, giving authors the opportunity to respond.

Consequences of non-disclosure

Undisclosed or prohibited AI use is a breach of publication ethics and is handled under our Publication Ethics framework and COPE guidance. Depending on severity and stage, outcomes may include:

• A request for clarification and a corrected disclosure statement.
• Rejection of a manuscript under review.
• A published correction where the issue is limited and the findings remain reliable.
• Retraction of the published article where AI misuse undermines the reliability of the work — for example, fabricated data, fabricated references, or manipulated images (see Corrections & Retractions).
• Notification of the authors' institution in serious cases.

A developing area

AI capabilities and community norms are evolving quickly. Lumora reviews this policy regularly against COPE and ICMJE guidance and will update it as standards develop. Questions about whether a specific use requires disclosure are welcome at Support@lumora.sa — when in doubt, disclose.

Last updated: July 2026 · Related: Publication Ethics, Peer Review Policy, Author Guidelines, All Policies