Generative AI Policy

INTRODUCTION

This policy is based on and refers to the guidelines outlined in the Generative AI Policies for Journals, as provided by:

Jurnal Fisika Unand (JFU) recognizes that generative artificial intelligence (AI) and AI-assisted technologies can support scientific research, manuscript preparation, scholarly communication, programming, modeling, simulation, data analysis, and visualization.

However, these technologies also introduce risks relating to scientific accuracy, fabricated information, bias, research integrity, reproducibility, confidentiality, privacy, copyright, intellectual property, and accountability.

This policy establishes requirements for authors, reviewers, editors, Editorial Board members, and journal personnel involved in the submission, evaluation, production, and publication of scholarly manuscripts.

The policy is informed by relevant guidance from the Committee on Publication Ethics (COPE), including its position on AI and authorship and its discussion of AI in editorial decision-making. It also considers relevant recommendations from the International Committee of Medical Journal Editors (ICMJE), where applicable to medical physics, medical imaging, human participants, and health-related research.

References to these organizations do not imply their endorsement or certification of JFU.

The fundamental principles of this policy are human accountability, transparency, scientific integrity, confidentiality, independent scholarly judgment, and responsible AI use.

1. Human Responsibility and Accountability

Human authors, reviewers, and editors remain responsible for work performed with the assistance of generative AI.

AI use does not reduce responsibility for scientific accuracy, originality, attribution, methodological soundness, research ethics, data integrity, confidentiality, copyright, reproducibility, or compliance with journal policies.

AI-generated material must be critically examined and independently verified by the responsible human contributor before use.

No person may transfer accountability for scientific content or editorial decisions to an AI system.

2. AI Tools and Authorship

Generative AI systems must not be listed as authors or co-authors.

AI systems cannot meet JFU's authorship requirements because they cannot approve manuscripts, assume accountability, disclose competing interests, enter into publication agreements, or respond responsibly to questions about research integrity.

Only eligible human contributors may be recognized as authors.

All listed authors must meet the requirements of JFU's Authorship and Contributorship Policy.

3. Acceptable Uses of AI by Authors

Authors may use generative AI when its use is lawful, scientifically justified, ethically appropriate, transparent, and subject to meaningful human oversight.

Potentially acceptable applications include:

  • Spelling, grammar, language, and readability assistance.

  • Translation of author-written material.

  • Literature searching and organization.

  • Software development, code debugging, and computational assistance.

  • Data processing and statistical analysis.

  • Mathematical modeling and numerical simulation.

  • Scientific visualization and graphical presentation.

  • Research ideation and preliminary methodological exploration.

Authors must verify all AI-assisted outputs against appropriate scientific evidence and reliable sources.

Permitted assistance must not compromise originality, scientific validity, reproducibility, confidentiality, intellectual property, or research integrity.

4. Prohibited Uses of AI by Authors

Authors must not use generative AI to:

  • Fabricate or falsify research data, measurements, experiments, observations, or results.

  • Generate fictitious references, quotations, DOIs, approvals, or funding declarations.

  • Conceal plagiarism, duplicate publication, or inappropriate text recycling.

  • Misrepresent synthetic or simulated data as real experimental observations.

  • Create deceptive scientific images or visualizations.

  • Conceal methodological limitations, uncertainty, or bias.

  • Replace essential scientific validation or independent judgment.

  • Process confidential or identifiable information without appropriate authorization and safeguards.

  • Violate copyright, privacy, consent, or licensing requirements.

  • Fabricate authorship, peer-review reports, or editorial records.

Serious or deceptive violations may result in rejection, correction, retraction, or other proportionate action under JFU's research integrity policies.

5. Disclosure of AI Use by Authors

Authors must disclose substantive uses of generative AI or AI-assisted technologies.

Where relevant, the disclosure must identify:

  • The name, provider, and version or access date of the tool.

  • The purpose of its use.

  • The manuscript section or research stage affected.

  • The nature of the input and output.

  • The verification or validation procedures performed.

  • Material limitations or risks.

AI used for manuscript preparation must be disclosed in a Generative AI Declaration.

AI used in research methods, data analysis, code development, modeling, simulation, image processing, or the generation of research materials must also be described in the Methods section with sufficient detail for scientific evaluation.

Transparent disclosure does not automatically disadvantage a manuscript, and disclosure alone does not establish that the use was appropriate.

6. Language and Copyediting Assistance

Disclosure is not normally required for routine spelling, punctuation, grammar, or formatting correction when the tool does not generate substantive scholarly content, alter scientific meaning, introduce references, or process confidential information inappropriately.

Disclosure is required when AI generates or substantially rewrites text, summarizes or interprets literature, translates substantive scholarly content, creates scientific claims, produces references, contributes to analysis, generates code used in research, or materially modifies scientific images.

Authors remain responsible for reviewing all language and copyediting changes.

7. Generative AI Declaration

Authors who have used generative AI substantively must include a declaration such as:

Declaration of AI use:

During the preparation of this work, the authors used [tool name, provider, version or access date] for [specific purpose]. The tool was used in [manuscript section or research stage]. The authors independently reviewed, verified, and revised the relevant outputs and accept full responsibility for the accuracy, originality, integrity, and final content of the manuscript.

When AI forms part of the research methodology, authors must provide additional technical information in the Methods section.

Where no substantive generative AI was used, authors may state:

Declaration of no substantive AI use:

The authors did not use generative AI or AI-assisted technologies to generate substantive scholarly content, data, analyses, images, or conclusions in this work.

8. AI as a Scholarly Source, Citations, and References

Generative AI systems are not authoritative scholarly sources.

AI-generated responses must not be treated or cited as independent scientific evidence, expert authority, or substitutes for identifiable and verifiable primary literature.

Authors must not rely on an AI-generated answer as the authoritative basis for a scientific claim, physical theory, experimental result, computational finding, clinical conclusion, or methodological recommendation.

When AI suggests publications, datasets, standards, or other references, authors must independently consult and verify the underlying sources before citation.

Authors are responsible for confirming that every cited source exists, that its bibliographic details and DOI are accurate, and that it supports the relevant statement.

Fabricated references, nonexistent publications, invented quotations, and misleading citations are prohibited.

Where an AI system is itself the subject of research, its outputs may be reported as research data or materials, provided that the system, version, methodology, settings, and limitations are documented appropriately. This does not make the AI system an authoritative scientific source.

9. AI-Generated and AI-Modified Scientific Images

AI-generated or materially AI-modified images, figures, graphical elements, and visualizations must be disclosed.

Authors must explain the tool used, its purpose, the elements generated or modified, the procedures used to evaluate scientific accuracy, and whether the image represents observed, simulated, reconstructed, enhanced, or synthetic information.

JFU permits scientifically justified AI-assisted reconstruction, segmentation, denoising, enhancement, simulation, and visualization when these processes are adequately described and validated.

AI must not be used to create false experimental or clinical images, conceal or relocate scientifically meaningful features, misrepresent physical measurements, or present synthetic images as direct observations.

Authors must retain original files and relevant processing records.

10. Synthetic, Simulated, and Augmented Data

Synthetic and simulated data may be used when scientifically justified, transparently identified, appropriately validated, and adequately described.

Authors must distinguish between:

  • Observed or measured data.

  • Simulated physical data.

  • Synthetic datasets.

  • Augmented training data.

  • Reconstructed data.

  • Synthetic images.

  • Benchmarking and validation data.

The manuscript must describe relevant models, inputs, assumptions, generation procedures, validation methods, limitations, potential biases, and the availability of supporting data or code.

Synthetic or simulated data must never be misrepresented as measurements or observations obtained from the physical world.

11. AI-Assisted Programming and Computational Research

Authors may use AI to assist with software code, algorithms, scripts, numerical modeling, simulations, and computational workflows.

Authors remain responsible for verifying code functionality, testing computational outputs, documenting software environments, identifying security and licensing risks, validating numerical results, and disclosing substantive AI contributions.

AI-generated code must not be represented as an exclusively human-created contribution when substantive AI assistance was involved.

Research reproducibility requirements remain applicable regardless of whether code was developed with AI assistance.

12. Confidentiality, Privacy, and Sensitive Information

Authors, reviewers, editors, and journal personnel must protect confidential, proprietary, restricted, identifiable, and sensitive information.

Such information must not be entered into AI systems unless the processing is appropriately authorized, legally permitted, consistent with relevant consent and ethics requirements, and supported by adequate security and confidentiality safeguards.

Particular care is required for patient data, medical images, unpublished manuscripts, confidential peer-review information, proprietary datasets, embargoed research, and security-sensitive material.

Users must minimize the information provided to AI systems and must not assume that an AI service provides adequate confidentiality merely because it is commercially available.

AI use must comply with JFU's Privacy Statement and Ethical Oversight of Research Policy.

13. Copyright and Intellectual Property

Authors are responsible for ensuring that AI-assisted content does not infringe copyright, violate licensing conditions, disclose protected information, or reproduce third-party material without authorization.

AI-generated output must not automatically be assumed to be original, accurate, copyright-free, or legally reusable.

Authors must ensure that the material included in their manuscripts can lawfully be published under the applicable JFU licensing arrangements.

14. Approval and Authorization of AI Systems

An AI system may be approved by JFU or an authorized institution for a specifically defined purpose after appropriate consideration of confidentiality, information security, personal-data protection, data retention, model training, intellectual property, access controls, and human oversight.

Approval for one activity does not authorize use for unrelated activities.

The absence of a published list of approved AI systems does not authorize editors or reviewers to use external AI services with confidential manuscript material.

Journal authorization must be explicit and limited to the approved purpose.

15. Use of AI by Reviewers

JFU requires reviewers to conduct independent scholarly assessments.

Reviewers must personally read, analyze, and evaluate assigned manuscripts.

Reviewers must not use generative AI to:

  • Evaluate manuscript originality, scientific validity, significance, or methodological quality.

  • Summarize, condense, or synthesize confidential manuscript content for review preparation.

  • Assess experimental designs, numerical models, results, or conclusions on their behalf.

  • Generate substantive criticisms, recommendations, or scientific judgments.

  • Produce an AI-generated peer-review report represented as the reviewer's own assessment.

  • Determine acceptance or rejection recommendations through automated evaluation.

Reviewers must not upload unpublished manuscripts, abstracts, figures, tables, supplementary materials, reviewer reports, author information, or editorial correspondence to unauthorized AI systems.

Limited spelling, grammar, punctuation, or readability assistance for a reviewer-authored report may be permitted only when the tool is authorized for that purpose, confidentiality is protected, and the assistance does not generate or modify substantive scholarly assessments.

Reviewers must disclose any specifically authorized AI assistance to the handling editor when required.

The reviewer remains fully responsible for every scientific comment, recommendation, and conclusion.

Reviewers who cannot complete an assessment without prohibited AI assistance must decline the assignment.

16. Use of AI by Editors and Journal Personnel

Editors and journal personnel must preserve manuscript confidentiality, protect reviewer identities, and maintain independent editorial judgment.

AI may support specifically authorized administrative or quality-control functions where confidentiality, security, fairness, transparency, and meaningful human oversight are maintained.

AI-generated assessments, classifications, scores, or recommendations must not independently determine:

  • Desk rejection.

  • Reviewer selection.

  • Manuscript acceptance or rejection.

  • Research misconduct findings.

  • Appeal outcomes.

  • Corrections or retractions.

  • Other consequential editorial decisions.

Automated outputs must be interpreted critically by a designated human editor.

Where an approved AI system materially contributes to manuscript assessment, JFU will disclose the nature and purpose of that use to the affected author, subject to legitimate confidentiality, security, and integrity-investigation restrictions.

Routine low-risk administrative automation does not ordinarily require manuscript-specific notification.

17. Disclosure of Journal-Operated AI and Automated Tools

JFU is committed to publicly disclosing relevant AI and automated tools used in its editorial and publishing operations.

The journal distinguishes routine workflow automation from tools that may materially influence scholarly assessment or editorial decisions.

The public disclosure will identify, where applicable:

  • The tool or system and its provider.

  • Its authorized function.

  • Whether unpublished manuscript content or personal information is processed.

  • Whether outputs contribute to manuscript screening or assessment.

  • The safeguards applied to confidentiality, security, and potential bias.

  • The nature of human oversight.

JFU will not claim that a tool is in operational use unless its use has been verified.

Current Editorial Technology Disclosure

System or Function Declared Purpose Operational Status
Open Journal Systems (OJS) Manuscript submission and editorial workflow management In use
Turnitin Manuscript similarity screening Described in the journal's APC information; operational details to be confirmed
Generative AI for manuscript assessment AI-assisted editorial evaluation [CONFIRM WHETHER USED OR NOT USED]
AI-generated-content detection Preliminary screening of possible AI-generated material [CONFIRM WHETHER USED OR NOT USED]
AI-assisted editorial copyediting or language processing Language or production assistance [CONFIRM WHETHER USED OR NOT USED]
Other automated editorial decision-support systems [IDENTIFY PURPOSE, IF APPLICABLE] [CONFIRM STATUS]

JFU will update this disclosure when relevant tools are introduced, discontinued, or materially changed.

No automated tool may replace the final responsibility of a qualified human editor.

18. AI-Detection and Automated Screening

JFU may use authorized AI-detection or forensic tools as preliminary screening aids.

An AI-generated-content probability score, detector label, change in writing style, or automated assessment does not independently establish misconduct.

JFU will not make adverse editorial decisions solely on the basis of an AI-detection result.

When concerns arise, editors will assess the available evidence, manuscript content, author disclosures, methodological records, and relevant explanations.

Authors will normally receive a reasonable opportunity to respond to credible concerns.

19. Undisclosed or Inappropriate AI Use

Potential violations include undisclosed substantive AI use, fabricated references, manipulated research data, misleading synthetic images, unauthorized processing of confidential material, AI-generated peer-review reports, and other uses that compromise scientific integrity.

Failure to disclose AI use does not automatically establish research misconduct.

JFU will consider the nature, materiality, consequences, and circumstances of an omission, including whether the conduct was deceptive, whether the scientific record was affected, and whether transparent correction is possible.

Minor or inadvertent omissions may be addressed through clarification or revision.

Serious or deceptive conduct may be referred for further investigation under JFU's publication misconduct procedures.

20. Supporting Information and Verification

JFU may request relevant prompts, inputs, outputs, code, model settings, validation records, image-processing documentation, or other supporting information when reasonably necessary for scientific or ethical assessment.

Requests must be proportionate and must respect privacy, confidentiality, intellectual property, and applicable legal restrictions.

Where full disclosure is inappropriate, alternative verification methods may include redacted documentation, methodological summaries, confidential institutional confirmation, or independent technical assessment.

21. Editorial Actions and Post-Publication Outcomes

Depending on the evidence and seriousness of a concern, JFU may:

  • Request clarification or an amended AI declaration.

  • Require additional methodological or validation information.

  • Request correction or replacement of problematic material.

  • Suspend editorial processing while a concern is assessed.

  • Obtain specialist advice.

  • Reject a manuscript.

  • Replace a reviewer.

  • Refer a matter to an appropriate institution or authority.

  • Publish a correction or expression of concern.

  • Retract an article when justified.

Actions must be proportionate and consistent with JFU's Publication Ethics, Allegations of Research and Publication Misconduct, and Corrections and Retractions policies.

22. Complaints and Appeals

Authors, reviewers, and other affected individuals may appeal decisions under this policy when they identify a material factual error, procedural irregularity, relevant overlooked evidence, unmanaged competing interest, or significant misclassification of AI use.

Disagreement with editorial judgment alone does not establish grounds for appeal.

Appeals are handled under JFU's Complaints and Appeals Policy.

23. Policy Review and Updates

JFU will periodically review this policy to reflect developments in generative AI technologies, scientific publishing, research integrity standards, data protection, and relevant ethical guidance.

Material revisions will be communicated through the journal's official website.

This policy operates alongside the journal's Author Guidelines, Publication Ethics, Peer Review Policy, Authorship and Contributorship Policy, Research Data and Reproducibility Policy, Privacy Statement, and other applicable editorial policies.