AI Use Policy

This policy applies to AI used in manuscript preparation, research communication, data analysis, image or figure preparation, peer review, editorial assessment, quality control, publication production, and the development of AI-assisted technologies operated or commissioned by the Journal. AI remains an assistive technology within the Journal's publishing framework: human Authors, Peer Reviewers, Editors, Guest Editors, Editorial Board members, and Student Advisory Board members retain responsibility for their respective contributions, assessments, and decisions. Authors are encouraged to weigh not only whether a use of AI is technically possible, but how it may affect the originality, credibility, interpretation, reproducibility, and long-term impact of their work.

• Human accountability: Authors remain responsible for the originality, accuracy, interpretation, and integrity of their work; Editors remain responsible for editorial decisions; Peer Reviewers remain responsible for their independent assessments.
• Human assessment: AI can support scholarly activity, while scientific conclusions, interpretations, editorial decisions, and peer-review assessments stay subject to appropriate human judgement.
• Transparency: appropriate disclosure of AI use helps Readers understand how a work was developed and supports confidence in the published record.
• Confidentiality and data protection: Authors, Reviewers, and Editors consider carefully what information they provide to an AI system and whether it offers appropriate safeguards for confidential, personal, unpublished, or proprietary material.
• Originality and attribution: AI-assisted material is used in a way that accurately reflects the human contribution and complies with the Journal's Authorship and Originality policies.
• Proportionate use: the appropriate level of care depends on the purpose of the AI use, the information provided to the system, the nature of its output, and how much the AI shapes the work.

The Journal offers three categories to help Authors think through AI use: Okay, Watch, and Stop. These are a shared language for discussing AI use. Most everyday uses fall within Okay.

Okay: routine assistive use

• Spelling, grammar, and language correction
• Improving clarity or readability
• Supporting accessibility and language-diversity needs
• Translation
• Formatting and structural suggestions
• Comparable assistance that stays subject to the Author's own review

Authors may use AI for these purposes where the underlying research, reasoning, and scholarly contribution remain their own, and may note the use if it helps show how the work came together.

Watch: use that may shape the intellectual content

• Suggesting analytical or methodological approaches
• Drafting or substantially restructuring explanatory passages
• Helping interpret statistical or computational output
• Comparing findings against existing literature
• Identifying patterns within exploratory analysis

The Author stays able to explain, verify, and stand behind the resulting work, and lets the Journal know where AI has meaningfully shaped it.

Stop: use that substantially involves AI authorship

• Generating substantial manuscript content without meaningful human contribution
• Presenting AI output as human analysis, reasoning, or interpretation
• Fabricating or materially altering data, results, or references
• Assigning authorship or accountability to an AI system
• Delegating the entire peer review to an AI system
• Sharing confidential manuscripts or peer-review material with an unsecured AI system
• Using AI to paraphrase another person's work then claiming ownership

The Journal conducts an initial manuscript assessment before peer review to confirm compliance with the minimum requirements of its Editorial policy. Where concerns arise regarding undeclared or potentially substantive AI involvement, the Journal may take appropriate steps to determine whether the manuscript is suitable to proceed to peer review. This may include editorial assessment of the manuscript for indicators of AI-generated content, the use of established plagiarism or originality-checking tools, and direct communication with the Authors to clarify the nature and extent of any AI use. The Journal does not use external AI services to determine AI authorship or attribution, as such use may introduce additional concerns relating to confidentiality, data protection, intellectual property, transparency, and the scope of this AI Use Policy.

Where a submission concerns the development, testing, modelling, optimisation, evaluation, or application of AI or machine-learning (ML) systems, AI outputs may form an essential part of the research evidence, methodology, or narrative, and are assessed as part of the study's methodology. Under this definition, Authors are allowed to represent direct AI outputs as part of the text, so long captioned appropropriately. Authors are expected to describe the relevant systems, methodologies, inputs, outputs, and validation procedures clearly, identify the exact model used and its service provider (where applicable), and remain responsible for the research design, interpretation, and reporting of those outputs.

An AI system cannot be listed as an Author of Journal content; authorship requires human accountability and is governed by the Journal's Authorship policy. Where AI assists with preparing a manuscript, human Authors remain responsible for verifying the accuracy, originality, integrity, and attribution of the resulting work, and should be able to demonstrate their own intellectual contribution where substantial AI-generated content is present. Fabricated or unverified references are treated as a serious defect: every reference must have been read by at least one Author and must resolve to the work it names.

The Journal supports responsible AI-assisted scholarship while holding clear expectations around originality and attribution. Where AI-assisted material raises questions of originality, the Journal may ask for evidence of the work's provenance, including drafts, underlying data, analysis, or code. AI detection tools, where used, form one part of a wider assessment alongside manuscript content, Author declarations, and supporting evidence. Where evidence indicates substantial non-original or fabricated material, the Journal applies its Plagiarism and Research Integrity policies.

Manuscripts under review may contain confidential, unpublished, or commercially sensitive material, so the Journal asks Reviewers to keep AI use within carefully considered and disclosed boundaries. Peer Reviewers retain personal responsibility for the expert assessment they submit; AI may support limited review-related tasks where the use is transparent and respects confidentiality, and any AI-generated suggestion stays subject to the Reviewer's own expert judgement. Reviewers must minimise the use of AI and may not delegate assessment to it, as set out in Our Reviewers.

A Reviewer who accepts an assignment undertakes the review personally, and does not hand off, delegate, or share the manuscript or review responsibility with a student, colleague, or other third party without the Journal's prior approval. Where a Reviewer believes assistance would be appropriate, they should contact the Journal before sharing any manuscript material.

The Journal does not use generative systems to make acceptance, rejection, or correction decisions. Editors remain accountable for editorial decisions; AI may support approved administrative, organisational, retrieval, or accessibility functions, while decisions on editorial suitability, peer review, acceptance, and publication remain matters of human editorial judgement. Any reader-facing intelligence the Journal adds in future, including reference-insight features, will be labelled at the point of use and will never alter the published record.

Readmy Reviewer is an optional, pre-peer-review stage built on Readmy AI's proprietary method (readmy.org), developed jointly with the Journal to quality-control submitted manuscripts before peer review. It produces a structural map of the manuscript without processing any confidential data. on the document, showing where markers concentrate. Those markers include:

• Claim markers: where assertion-type language appears and in which section, surfacing argumentative density across the manuscript
• Data mentions: where empirical and numeric statements sit, separating evidence-bearing sentences from interpretive ones
• Citation tracking: where a reference first appears, where it recurs, and whether it resolves to the reference list
• Section role mapping: what function a passage serves in the argument, independent of the heading an Author chose
• Further markers, including claim-to-support linkage, hedging and modality shifts, terminology and acronym first use, self-citation density, figure and table sequencing, numbering integrity, text anomalies, placeholder language, generated-language patterns, section length ratios, reference density, cross-module consistency, and near-duplicate paragraphs.

Authors submitting to the Journal are asked whether they consent to their manuscript being triaged by Readmy. Where an Author consents, Reviewers will be given access to Readmy, where they can load the paper to create a structural map of the content. No raw manuscript text is processed to produce an outcome, and Readmy does not itself review, score, or make any determination about the manuscript.

Readmy does not replace dedicated, human reviewers, and it does not automate the review process. It gives Reviewers and the editorial office a way to move through the structural parts and syntax of a manuscript faster, so their expertise goes toward assessing contribution, value, and impact. Readmy Reviewer is being developed alongside Readmy Insights and Papers Viewer, a companion tool to help visualise papers for accessibility needs, as part of a joint programme between Readmy and the Journal to make research more accessible and to disseminate science responsibly and ethically. The Readmy Insights functionality is built-in for JQBio.

AI use in creating, reconstructing, enhancing, or altering original research figures, images, or graphical data is subject to the Journal's research integrity requirements. Limited technical assistance, such as labelling, annotation, or accessibility formatting that leaves the underlying raw data and research representation unchanged, sits comfortably within Okay for routine use. Where AI-generated or AI-modified visual material is itself an essential part of AI or machine-learning research, Authors must describe it explicitly in the manuscript with enough methodological detail to establish its role.

Every submission carries an AI declaration. Where a tool was used beyond routine, non-substantive assistance, Authors name the tool and version, state what it was used for, describe what was incorporated into the work, and confirm who verified the output; where no tool was used, the declaration says so. The declaration is published alongside the work and is not held against a submission: an undeclared use discovered later is treated far more seriously than an honest disclosure.

Authors are encouraged to consider an AI system's data-handling practices before sharing research materials, manuscript content, datasets, code, personal information, or intellectual property with it, taking particular care with unpublished research, commercially sensitive material, proprietary methods, and participant information. Any such use should remain compatible with applicable consent, confidentiality, licensing, contractual, and institutional requirements.

Where declared, suspected, or identified AI use raises questions of authorship, originality, data integrity, confidentiality, or editorial accountability, the Journal may carry out an AI risk assessment, drawing on the circumstances, available evidence, Author declarations, and manuscript content, and may invite the Author, Reviewer, or Editor to provide clarification. Where an investigation is needed, the Journal may pause the relevant process while the matter is reviewed; Authors should anticipate that this may affect expected timelines. Early conversation is welcome where anyone is uncertain how a proposed use of AI fits within this policy.

The Journal reviews its approach to AI periodically as technologies, research practices, and regulatory expectations develop, and may introduce or adjust AI-assisted tools accordingly, provided their use stays consistent with this policy and with the Journal's obligations to Authors, Reviewers, Editors, and the integrity of the scholarly record.