1. Principles of Research Integrity

All research submitted to Health Nexus must comply with established principles of responsible scientific conduct, regardless of whether artificial intelligence has been used.

The journal's principal requirements include:

• Honesty and Accuracy: Research data, observations, results, interpretations, and scholarly claims must be reported faithfully and without fabrication or falsification.
• Methodological Transparency: Research procedures, analytical methods, technological assistance, and relevant limitations must be documented sufficiently for appropriate scholarly evaluation.
• Human Accountability: Authors, reviewers, and editors remain responsible for the accuracy, integrity, and consequences of their scholarly activities and decisions.
• Research Ethics: AI-assisted activities must comply with applicable ethical approvals, informed consent requirements, privacy protections, and professional standards.
• Intellectual Property: The use of AI must respect copyright, source attribution, licensing conditions, confidential information, and legitimate third-party rights.
• Scientific Reproducibility: Research involving AI must provide appropriate information about tools, methods, datasets, computational procedures, and validation where relevant.
• Proportionate Oversight: AI-related risks and integrity concerns must be assessed according to reliable evidence, the nature of the activity, and its potential effect on research or patient safety.

These principles apply across conventional medicine, Traditional Chinese Medicine (TCM), biomedical sciences, clinical and translational research, interdisciplinary health sciences, and public health.

2. Acceptable Use of AI Tools

Health Nexus permits responsible use of AI-assisted technologies when their application is scientifically appropriate, ethically justified, and consistent with the journal's disclosure and research integrity requirements.

2.1 Language and Manuscript Preparation

Permitted activities may include:

• Spelling, grammar, punctuation, and language correction.
• Improving the clarity, readability, and organization of author-created text.
• Translation, subject to complete human review and verification.
• Formatting assistance and preparation of documentation.
• Assistance with the organization of references or literature search results.
• Generation of preliminary drafting suggestions that are critically reviewed, revised, and verified by the authors.

AI-assisted language editing must not alter the scientific meaning of the manuscript, introduce unsupported statements, create nonexistent references, or misrepresent the original research.

2.2 Research and Analytical Assistance

AI-assisted technologies may also support research activities such as:

• Exploratory literature searching and organization.
• Code development, debugging, and computational workflow assistance.
• Data preprocessing, classification, annotation, and analysis.
• Statistical modelling, machine learning, and predictive analysis.
• Medical image analysis, signal processing, and clinically relevant computational methods.
• Development or evaluation of research instruments and analytical frameworks.
• Generation of appropriately labelled synthetic data or simulations for legitimate methodological research.

When AI forms part of the research methodology, its use must be described with sufficient detail to permit scientific assessment. Authors must distinguish between AI-supported research procedures and AI assistance used solely in manuscript preparation.

2.3 Conditions for Permitted Use

3. Prohibited Use of AI

Health Nexus prohibits the use of AI in ways that compromise research integrity, misrepresent scientific evidence, violate confidentiality, or undermine responsible academic and clinical judgment.

Prohibited practices include:

• False Authorship: Listing an AI system, chatbot, language model, or automated software agent as an author or coauthor.
• Data Fabrication: Generating or altering fabricated experimental observations, clinical records, participant responses, research measurements, or findings and presenting them as authentic evidence.
• Falsified Images: Creating or manipulating clinical, radiological, pathological, microscopic, or other scientific evidentiary images in a manner that misrepresents actual observations.
• Fabricated References: Submitting invented publications, nonexistent DOIs, false quotations, or citations that materially misrepresent the referenced source.
• Misrepresentation of Synthetic Data: Presenting generated datasets, simulated patients, or artificial observations as genuine human, clinical, or experimental records.
• Undisclosed Substantive AI Use: Concealing material AI-generated or AI-assisted research, analysis, interpretation, or manuscript preparation in violation of the journal's disclosure requirements.
• Confidentiality Violations: Uploading protected clinical information, confidential research records, or unpublished manuscript materials to AI services without appropriate authorization and safeguards.
• Unverified Scientific Content: Presenting AI-generated findings, interpretations, statistical results, or clinical recommendations without adequate human evaluation.
• Misleading Attribution: Using AI-assisted material in a manner that infringes intellectual property rights or obscures the contribution of original sources.
• Manipulation of Peer Review: Using AI to fabricate reviewer reports, impersonate reviewers, misrepresent expert evaluation, or improperly influence editorial decisions.

These restrictions do not prohibit legitimate research involving synthetic data, AI-generated representations, computational image reconstruction, or AI-assisted clinical models when the procedures are scientifically justified, appropriately disclosed, ethically approved where necessary, and accurately represented.

4. Disclosure Requirements

Authors must disclose relevant use of generative AI and AI-assisted technologies in manuscript preparation, research design, analysis, interpretation, or scientific content generation. Disclosure is intended to support transparency, accountability, and informed editorial assessment.

4.1 Information to Be Disclosed

An AI-use declaration should identify, as appropriate:

• The name of the AI tool or system.
• The developer or service provider.
• The model or software version, where available.
• The purpose and extent of AI assistance.
• The manuscript sections, research procedures, data analyses, images, or other materials affected.
• Relevant dates, prompts, configurations, or procedures where needed for methodological evaluation.
• The human review, verification, validation, and revision measures undertaken by the authors.

4.2 Location of the Disclosure

AI use should be documented according to its function:

• Cover Letter: Notify the Editorial Office of relevant AI-assisted manuscript preparation or research activities.
• AI-Use Declaration or Acknowledgments: Describe material AI assistance with writing, translation, language editing, or content preparation.
• Methods Section: Report AI used for research design, data collection, statistical analysis, image processing, coding, modelling, or scientific evaluation.
• Figure Legends: Identify AI-generated, synthetic, reconstructed, or materially AI-modified figures where relevant.
• Supporting Documentation: Provide relevant model descriptions, prompts, code, configurations, or validation materials where necessary and legally shareable.

During double-blind review, authors must preserve appropriate anonymization in files sent to external reviewers. Complete identifying declarations may be provided separately to the Editorial Office, while methodologically necessary AI information must remain sufficiently transparent for scholarly assessment.

4.3 Illustrative AI-Use Declarations

4.4 Routine Editing and Material AI Assistance

Conventional non-generative spelling correction, basic punctuation checking, and routine formatting functions that do not generate substantive content or alter scientific meaning do not ordinarily require a separate AI-use declaration.

Use of generative AI for substantive drafting, rewriting, translation, analysis, coding, or scientific content preparation must be disclosed. Where authors are uncertain whether an activity requires disclosure, they should identify the tool and purpose in the cover letter and seek editorial guidance.

Appropriate and transparent AI assistance is not itself grounds for rejection. Material nondisclosure, misleading reporting, or the introduction of scientifically unreliable content may require correction, clarification, or further research integrity assessment.

5. Authorship Responsibility

Artificial intelligence systems do not meet the journal's authorship requirements and must not be listed as authors, coauthors, corresponding authors, or individuals capable of assuming scholarly accountability.

All human authors are responsible for:

• Content Accuracy: Verifying factual statements, research findings, quotations, tables, figures, and supporting arguments.
• Reference Accuracy: Confirming that cited publications exist and support the claims for which they are cited.
• Data and Analysis: Ensuring the validity of reported data, statistical analysis, computational procedures, and scientific interpretation.
• Ethical Compliance: Maintaining appropriate research approvals, confidentiality, informed consent, and legal compliance.
• Originality: Avoiding plagiarism, inappropriate reuse, fabricated material, and intellectual property infringement.
• Final Approval: Reviewing and approving the submitted and published versions of the manuscript.
• Accountability: Responding to legitimate scholarly questions, editorial inquiries, corrections, and research integrity concerns.

AI assistance does not qualify an individual for authorship. Conversely, human authorship requires a genuine scholarly contribution and satisfaction of the journal's applicable authorship criteria.

Authors cannot transfer responsibility for errors, fabricated references, inappropriate image manipulation, methodological weaknesses, or unsupported conclusions to the AI tools they used.

6. AI in Peer Review

Manuscripts submitted to Health Nexus are confidential scholarly communications. Reviewers, editors, and editorial personnel must protect unpublished research, manuscript files, reviewer reports, author information, and related correspondence.

6.1 Responsibilities of Reviewers

Reviewers must not independently upload, disclose, or transfer confidential manuscript content to public or external AI systems where confidentiality and information security cannot be assured.

The following rules apply:

• Prior written authorization from the Editorial Office is required before using AI to assist with manuscript review.
• Any external processing of confidential manuscript content must also meet applicable author-permission, privacy, confidentiality, and data-protection requirements.
• Reviewers must identify the proposed AI tool and its intended purpose when requesting authorization.
• Authorized AI use must be disclosed to the Editorial Office.
• Reviewers must independently verify any AI-assisted analysis, summary, language revision, or observation.
• AI must not be used to infer blinded author identities or disclose confidential reviewer information.
• Reviewers remain responsible for the originality, accuracy, fairness, and scholarly substance of their reports.

AI-generated evaluations must not replace independent expert review. A reviewer may not delegate substantive scholarly judgment to an AI system.

6.2 Responsibilities of Editors

Editors must protect manuscript confidentiality and ensure that AI-assisted editorial activities are appropriately authorized, secure, and subject to human oversight.

• Confidential manuscript materials must not be uploaded to external AI services without appropriate authorization and safeguards.
• AI must not autonomously determine manuscript acceptance, rejection, or revision outcomes.
• AI must not replace independent editorial assessment of reviewer reports, scientific validity, or research integrity concerns.
• Automated systems must not independently decide allegations of misconduct, authorship disputes, or post-publication corrective actions.
• The Editor-in-Chief or an authorized delegated editor remains responsible for formal editorial decisions.

6.3 Use of AI by the Journal and Publisher

The journal or publisher may use suitable automated tools for limited administrative or technical functions, including similarity screening, reference verification, metadata processing, language assistance, and production quality checks, provided that appropriate confidentiality, accuracy, and oversight arrangements are maintained.

Where AI is used materially in editorial production or journal-generated content, the journal remains responsible for the accuracy of the resulting material and for appropriate disclosure of such use.

7. Data and Image Integrity

Scientific data, images, clinical records, and related research outputs must accurately represent the observations, evidence, and procedures from which they were derived.

7.1 Original Research Data

• Research data must not be fabricated, falsified, selectively altered, or misleadingly presented.
• AI-assisted data processing must be methodologically appropriate and documented where it affects the reported findings.
• Authors must preserve relevant original records, analytical procedures, and supporting documentation according to applicable requirements.
• Uncertainty, missing data, computational limitations, and material analytical decisions must be reported transparently.

7.2 Scientific and Clinical Images

AI-generated or AI-modified images must not be presented as authentic clinical, laboratory, pathological, radiological, microscopic, or other experimental evidence when they do not faithfully represent the underlying observations.

Where AI-assisted imaging forms a legitimate part of the research method, authors must:

• Explain the purpose of image generation, enhancement, segmentation, reconstruction, or modification.
• Identify relevant software, models, versions, and processing steps.
• Distinguish original source images from generated, reconstructed, or modified outputs.
• Retain source images and verification materials where available and appropriate.
• Identify material processing in the Methods section and relevant figure legends.
• Ensure that published images do not compromise patient confidentiality or applicable third-party rights.

7.3 Synthetic Data and Illustrative Content

Synthetic datasets and AI-generated illustrative content may be used where scientifically or educationally justified. Such materials must be clearly identified and must not be represented as real patient data, experimental observations, or independently verified clinical evidence.

Studies involving synthetic datasets should report relevant generation procedures, source-data characteristics, intended uses, validation measures, limitations, and privacy or re-identification risks where applicable.

7.4 Research Involving Medical AI

Manuscripts developing or evaluating AI systems for medical or health-related applications must report sufficient methodological information to permit assessment of validity, reliability, and clinical relevance.

Where applicable, authors should describe:

• The intended use and research purpose of the AI system.
• The model, developer, version, software environment, and relevant configuration.
• Dataset sources, participant or sample characteristics, inclusion criteria, and preprocessing methods.
• Training, validation, and test-set separation, including measures to prevent data leakage.
• Appropriate performance metrics, uncertainty estimates, and validation procedures.
• Potential bias, limitations, generalizability, and subgroup performance where relevant.
• The degree of human oversight and potential consequences of erroneous outputs.
• Data, code, model, and documentation availability or applicable access restrictions.

Findings from exploratory or retrospective AI evaluations must not be presented as proof of clinical safety, regulatory approval, treatment efficacy, or implementation readiness without sufficient supporting evidence. Research involving human participants or protected health information remains subject to applicable ethics review and data protection requirements.

8. Plagiarism and AI-Generated Content

Health Nexus applies originality and research integrity screening to submitted manuscripts. AI-generated or AI-assisted content is subject to the same standards of accuracy, originality, attribution, and scholarly responsibility as other manuscript content.

8.1 Similarity Screening

8.2 Reference and Citation Verification

Authors are responsible for verifying citations generated or formatted with AI assistance. They must:

• Confirm that referenced publications and datasets exist.
• Verify titles, authors, publication details, and DOI information.
• Confirm that cited sources substantively support the associated claims.
• Identify relevant corrections, withdrawals, or retractions where necessary.
• Follow the journal's AMA Citation Style requirements.

8.3 Automated AI-Content Detection

The journal may use appropriate automated tools to assist with preliminary screening for textual similarity, image irregularities, or potential AI-related concerns. Automated detection tools have limitations, including the possibility of false-positive and false-negative results.

An AI-detection score alone is not conclusive evidence of undisclosed AI use, plagiarism, data fabrication, or research misconduct. Editors must consider the relevant manuscript content, supporting documentation, author explanations, and other credible evidence before taking action.

A manuscript will not be rejected solely because an automated tool assigns it a high probability of AI-generated text. Substantive concerns concerning accuracy, authorship, originality, disclosure, or research reliability remain subject to independent editorial assessment.

9. Ethical Oversight and Misconduct

Suspected inappropriate use of AI is evaluated according to the journal's Publication Ethics Policy and applicable principles of procedural fairness, confidentiality, proportionality, and research integrity.

9.1 Preliminary Assessment

When an AI-related concern arises, the Editorial Office may:

1. Conduct a preliminary confidential assessment of the reported concern.
2. Identify the specific manuscript content, research procedure, or disclosure at issue.
3. Request clarification from the corresponding author and, where necessary, the coauthors.
4. Request relevant documentation, including source files, original images, analytical code, AI-use records, or methodological explanations.
5. Seek independent methodological, technical, statistical, clinical, or ethical advice where warranted.
6. Determine whether the concern requires correction, additional review, further investigation, or another proportionate editorial response.

9.2 Potential Editorial Actions

Depending on the circumstances and established evidence, actions may include:

• Requesting clarification or an amended AI-use disclosure.
• Returning the manuscript for technical or substantive revision.
• Requesting independent verification of data, images, analytical procedures, or results.
• Suspending peer review or production while a material concern is investigated.
• Rejecting a submission or withdrawing acceptance when justified.
• Publishing a correction, expression of concern, or retraction where appropriate.
• Notifying relevant institutions, ethics committees, or other appropriate authorities when supported by the circumstances and applicable procedures.

9.3 Proportionality and Author Response

The journal distinguishes minor administrative disclosure omissions from scientifically consequential errors, deliberate misrepresentation, falsification, confidentiality breaches, and other serious concerns.

Authors should normally be given an appropriate opportunity to explain material concerns and provide relevant evidence before a final determination is made. An allegation of misconduct is not treated as established solely on the basis of an automated detection result or an unsupported complaint.

9.4 Post-Publication Corrections and Appeals

Authors must notify the Editorial Office when they identify a material error attributable to AI-assisted research or manuscript preparation. Necessary corrections, expressions of concern, and retractions are addressed under the journal's publication ethics procedures and are not contingent on payment of an administrative service fee.

Authors may seek reconsideration of an AI-related editorial decision through the journal's Complaints & Appeals Policy.

10. Transparency and Accountability

Transparent reporting of AI-assisted scholarly activities is essential for evaluating scientific validity, research reproducibility, ethical compliance, and the reliability of published findings.

10.1 Methodological Transparency

Research involving AI should report, where relevant:

• The intended research purpose and function of the AI system.
• The AI model, provider, version, and evaluation period.
• Relevant computational methods, prompts, parameters, and analytical workflows.
• Research data sources, preprocessing procedures, and evaluation criteria.
• Validation procedures, performance limitations, uncertainty, and error analysis.
• Human verification, oversight, and responsibility for scientific or clinical decisions.
• Data, code, model, and documentation availability or applicable restrictions.

10.2 Data and Code Availability

All submissions must include an appropriate Data Availability Statement in accordance with the journal's Data Policy. For AI-related research, authors should additionally identify the availability of relevant datasets, scripts, model documentation, computational environments, and reproducibility materials where applicable.

The journal encourages responsible sharing but does not require unlawful disclosure of identifiable patient information, restricted clinical records, proprietary systems, or other protected materials.

10.3 Clinical Safety, Fairness, and Privacy

Studies involving AI applications in diagnosis, clinical prediction, treatment support, health monitoring, or other healthcare activities must address relevant risks to patient safety, privacy, and the reliability of medical judgment.

Where relevant, authors should consider:

• The clinical or research setting in which the system is intended to operate.
• Potential consequences of incorrect, incomplete, or misleading predictions.
• Performance across relevant populations or patient groups.
• Human oversight, escalation procedures, and opportunities to challenge AI outputs.
• Protection of personal health information and relevant consent conditions.
• Limitations affecting applicability, generalizability, or clinical implementation.

AI-assisted research is subject to the same foundational standards of scientific evidence and human-subject protection as research conducted without AI. Claims concerning clinical effectiveness, fairness, safety, or readiness for practice must be supported by appropriate evidence.

11. Policy Evolution and International Guidance

Artificial intelligence technologies and their applications in research and publishing continue to evolve. Health Nexus may periodically review and amend this policy to reflect developments in medical AI, scholarly publishing, research integrity, data protection, copyright, and applicable clinical or regulatory requirements.

11.1 Relevant International Guidance

This policy is informed by recognized guidance and scholarly standards, including:

• International Committee of Medical Journal Editors (ICMJE): Recommendations concerning AI-assisted scientific publishing, authorship accountability, disclosure, manuscript confidentiality, and editorial responsibilities. Official ICMJE Guidance
• Committee on Publication Ethics (COPE): Guidance concerning editorial accountability, research integrity, publication misconduct, corrections, and fair handling of ethical concerns. COPE
• EQUATOR Network: Relevant reporting guidance supporting transparent and complete reporting of medical research, including studies using artificial intelligence. EQUATOR Network
• CONSORT-AI and SPIRIT-AI: Reporting guidance relevant to clinical trials involving AI interventions and their protocols.
• TRIPOD+AI: Reporting guidance relevant to the development and evaluation of clinical prediction models.
• Other Applicable Guidance: Appropriate standards for AI diagnostic accuracy, medical imaging, research data governance, human-subject protection, and responsible computational research.

Authors must follow reporting guidance appropriate to their study design. Citation of a guideline or inclusion of a checklist does not independently establish methodological validity, ethical approval, or compliance with applicable law.

11.2 Policy Consistency

This policy should be read together with the journal's Author Guidelines, Data Policy, Peer Review Process, Publication Ethics Policy, Copyright & Licensing Policy, and applicable reporting requirements.

References to external organizations and scholarly frameworks indicate the guidance informing the journal's policy. They do not independently establish formal membership, certification, endorsement, or database indexing.

11.3 Responsibilities Following Policy Updates

Authors, reviewers, and editors are responsible for consulting the current version of this policy and complying with applicable ethical, institutional, contractual, legal, and regulatory requirements. Material policy amendments should be communicated clearly, including any relevant effective dates or transitional arrangements.