1. Purpose and Scope

This policy establishes the conditions under which generative AI and AI-assisted technologies may be used in research, manuscript preparation, peer review, editorial work, and publication activities associated with IJoDS.

For this policy, AI tools include generative AI systems, LLMs, multimodal models, AI agents, AI-assisted coding systems, machine-learning services, image-generation systems, automated analytical systems, and related technologies capable of generating, modifying, selecting, summarizing, interpreting, translating, analyzing, or transforming scholarly content.

Because IJoDS publishes research in machine learning, deep learning, natural language processing, predictive modelling, computational simulation, big-data analytics, statistical analysis, software development, and related areas, this policy distinguishes between:

• AI used to prepare or improve a manuscript; and
• AI used as an object, instrument, method, model, algorithm, or computational component of the research itself.

These uses have different reporting requirements. AI used in the research must be described as part of the research methodology. AI used to prepare the manuscript must be disclosed according to Section 6 of this policy.

2. Human Responsibility and Accountability

AI tools must not replace human critical thinking, disciplinary expertise, methodological judgment, interpretation, or editorial responsibility. Every person who uses an AI tool in connection with IJoDS remains responsible for the resulting work.

Authors are responsible for the accuracy, originality, validity, integrity, and interpretation of their manuscript, including material produced or modified with AI assistance. In particular, authors must:

• review and verify AI-generated output before using it;
• correct factual errors, unsupported claims, misleading statements, and biased outputs;
• independently verify citations, quotations, numerical values, statistical interpretations, and references;
• ensure that the manuscript reflects the authors' own scientific reasoning and conclusions; and
• ensure compliance with confidentiality, privacy, copyright, licensing, and research-ethics requirements.

3. Use by Authors in Manuscript Preparation

Authors may use AI tools in a limited and supportive role during manuscript preparation, provided that human oversight is maintained throughout the process.

Appropriate uses may include:

• improving language, readability, grammar, and clarity;
• reorganizing author-written material for clearer presentation;
• assisting with translation, provided that a competent human verifies the translation;
• helping identify keywords, concepts, or relevant areas for further literature searching;
• supporting formatting or checks against IJoDS manuscript requirements;
• assisting with code development or debugging when this use is transparently reported as part of the research method where applicable; and
• supporting the preparation of permitted explanatory diagrams or other non-evidentiary visual material.

AI-generated material must not be inserted into a manuscript without critical human review. The final manuscript must represent an authentic scholarly contribution by the named human authors.

4. Prohibited and Unacceptable Uses

AI tools must not be used to compromise research integrity or misrepresent scholarly work. Prohibited uses include, but are not limited to:

• fabricating, falsifying, or deceptively altering research data or results;
• inventing participants, observations, experiments, interviews, survey responses, datasets, or empirical evidence;
• creating fabricated references, citations, quotations, DOIs, sources, or bibliographic information;
• presenting AI-generated interpretations as if they were independently developed and verified scholarly judgments of the authors;
• generating substantial scientific claims or manuscript sections without genuine human intellectual contribution and verification;
• concealing plagiarism, redundant publication, inappropriate text reuse, or copyright infringement;
• using generative AI to create or manipulate primary research images in a manner that changes their evidentiary meaning;
• using AI to circumvent ethical approval, informed-consent requirements, data-protection obligations, or research-integrity requirements; or
• uploading confidential or protected third-party material to an AI service when the author does not have the right to do so.

Suspected fabrication, falsification, plagiarism, deceptive image manipulation, fabricated citations, or other serious misuse may be handled under the IJoDS Allegations of Misconduct policy.

5. AI and Authorship

AI tools, LLMs, chatbots, AI agents, and other automated systems must not be listed as authors or co-authors of an IJoDS article and must not be cited as authors.

Authorship requires responsibilities that can only be fulfilled by humans, including making a substantial intellectual contribution, critically drafting or revising the work, approving the final version, consenting to submission and publication, disclosing conflicts of interest, and accepting accountability for the accuracy and integrity of the work.

The use of AI does not reduce or transfer any author's responsibilities under the IJoDS Authorship and Contributorship policy.

6. Disclosure of AI Use in Manuscript Preparation

Authors must disclose material use of generative AI or AI-assisted technologies in manuscript preparation. The declaration must be sufficiently specific to allow editors, reviewers, and readers to understand how AI contributed to the submitted work.

The declaration should identify, where relevant:

• the name of the AI tool or service;
• the model or version, when known and relevant;
• the purpose for which it was used;
• the part or stage of manuscript preparation in which it was used; and
• confirmation that the authors reviewed, verified, and revised the output and accept responsibility for the final manuscript.

Recommended heading:

Declaration of Generative AI and AI-Assisted Technologies in Manuscript Preparation

Recommended statement:

During the preparation of this manuscript, the author(s) used [NAME OF TOOL/SERVICE AND VERSION, IF RELEVANT] for [PURPOSE]. The author(s) subsequently reviewed, verified, and revised the affected material as necessary and take full responsibility for the accuracy, integrity, and content of the final manuscript.

Disclosure is a transparency requirement and, by itself, is not grounds for rejection when the AI use complies with this policy.

Uses that normally do not require disclosure

Disclosure is not normally required when tools are used only for:

• basic spelling, punctuation, or conventional grammar checking that does not materially rewrite the text;
• standard word-processing functions;
• conventional reference-manager functions that do not use generative AI to select, create, summarize, or modify scholarly content; or
• specialist accessibility technologies used solely to facilitate accessibility.

If a tool materially rewrites, generates, reorganizes, summarizes, translates, selects, or interprets scholarly content, its use should be disclosed.

7. AI as Part of the Research Method

When AI, machine learning, an LLM, an AI agent, an external model API, or another automated system is part of the research itself, its use must be reported in the Materials and Methods section in accordance with the IJoDS Author Guidelines.

The description must be sufficiently detailed for readers to understand the procedure and, where reasonably possible, reproduce or independently evaluate it. Authors should report the following when relevant:

• model, algorithm, tool, service, or API name;
• developer or provider;
• model and software version, model identifier, or access date when a version is not fixed;
• research purpose and role of the AI system;
• input data and relevant preprocessing steps;
• training, fine-tuning, retrieval, or adaptation procedures;
• important parameters and configuration settings;
• prompting strategy or prompt templates when they materially affect the results;
• random seeds or sampling settings when available and relevant;
• human review, validation, benchmarking, or error-checking procedures;
• evaluation datasets and performance metrics; and
• known limitations that may affect interpretation or reproducibility.

AI used as a research method is not adequately disclosed by an AI manuscript-preparation declaration alone. Methodological AI use must be described in Materials and Methods.

8. Code, Models, Prompts, APIs, and Computational Reproducibility

Computational reproducibility is particularly important for IJoDS. When AI-assisted systems are used to generate, modify, debug, translate, or explain research code, authors remain responsible for the validity, security, licensing status, and reproducibility of the final code.

Where relevant to reproduction or verification, authors should preserve and make available:

• source code or executable workflows;
• dependency and package versions;
• model checkpoints or model identifiers when redistribution is permitted;
• configuration files;
• prompt templates, system instructions, or structured inputs that materially affect outputs;
• API model identifiers and relevant access dates;
• random seeds and sampling parameters where supported; and
• evaluation scripts and procedures.

Availability of code, data, and computational materials should follow the IJoDS Data and Reproducibility policy. If proprietary systems or closed APIs prevent full reproducibility, authors must state the limitation clearly rather than imply that the procedure is fully reproducible.

9. Synthetic Data and AI-Generated Research Data

Synthetic data may be used when scientifically justified, but it must never be presented as directly observed empirical data. Authors must clearly distinguish synthetic, simulated, augmented, imputed, and observed data.

When synthetic data materially contribute to a study, authors should report:

• why synthetic data were used;
• the method, model, or system used to generate them;
• the proportion or role of synthetic data relative to observed data;
• relevant generation settings and validation procedures;
• potential biases, privacy risks, or distributional differences; and
• how the use of synthetic data affects the interpretation and generalizability of the results.

Synthetic data derived from personal, confidential, health, or otherwise sensitive information remain subject to the IJoDS Ethical Oversight and Privacy Statement.

10. Literature Searching, Reviews, Citations, and References

AI tools may assist authors in exploring literature, identifying search terms, organizing references, or screening material, but AI output must not be treated as an authoritative bibliographic source.

Authors are responsible for every citation and reference in the manuscript and must independently verify that:

• the cited work exists;
• the title, authors, source, year, pages, and DOI or other identifier are correct;
• the source actually supports the statement for which it is cited;
• quotations accurately reproduce the original source; and
• references comply with the IJoDS IEEE reference style.

AI-generated or AI-suggested references must not be accepted without verification against the original source. Fabricated or materially inaccurate references may result in rejection or investigation under the IJoDS Allegations of Misconduct policy.

For systematic literature reviews, systematic mapping studies, bibliometric studies, and other evidence-synthesis methods, the documented search, selection, and analysis procedures must remain transparent and reproducible. AI-assisted screening or classification must be reported when it materially affects study selection or analysis.

11. Figures, Images, and Data Visualizations

11.1 Explanatory and conceptual figures

AI tools may assist in creating explanatory figures such as conceptual diagrams, flowcharts, decision trees, timelines, schematic illustrations, or workflow diagrams when the figure represents the authors' own scholarly interpretation and does not purport to be primary research evidence.

Authors must verify the scientific accuracy and originality of the figure, ensure that its use does not infringe third-party rights, and disclose material generative-AI assistance in the figure caption as well as in the general AI declaration.

11.2 Data visualizations

AI-assisted tools may be used to support plots, charts, graphs, heatmaps, and other data visualizations only when the output is faithfully and reproducibly derived from genuine underlying data. AI must not fabricate, invent, alter, or conceal the underlying data or results.

When AI is part of the analytical or visualization workflow, the relevant tool, model, version, and method must be described in Materials and Methods. The underlying data remain subject to the IJoDS Data and Reproducibility policy.

11.3 Primary research images

Generative AI must not be used to create, reconstruct, fabricate, or deceptively alter primary research images or observed evidence. AI must not add, remove, replace, obscure, or invent features in images in a way that changes their evidentiary meaning.

Legitimate AI-based image analysis used as part of the scientific method is permitted when it is methodologically justified, transparently described, and reproducible.

11.4 Graphical abstracts and promotional artwork

If IJoDS requests or accepts a graphical abstract, authors should use scientific or professional illustration tools with licensing terms that permit scholarly publication. General-purpose generative image systems should not be used to create graphical abstracts. AI-generated cover or promotional artwork requires prior approval from the Editor-in-Chief and publisher.

12. Privacy, Confidentiality, and Intellectual Property

Before submitting any material to an AI service, users must consider the service's privacy policy, terms of use, data-retention practices, model-training practices, and rights granted over uploaded material and generated output.

Authors must not upload confidential, proprietary, unpublished, personally identifiable, or otherwise protected information to an AI service unless they have the legal and ethical right to do so and the service provides appropriate safeguards.

Particular care is required for:

• personal or identifiable participant information;
• health, clinical, behavioral, or other sensitive data;
• restricted or unpublished datasets;
• confidential third-party documents;
• copyrighted material;
• proprietary source code or research information; and
• manuscripts and communications received through peer review.

These requirements operate together with the IJoDS Privacy Statement, Intellectual Property policy, and Ethical Oversight policy.

13. Use of AI by Peer Reviewers

IJoDS uses a double-blind peer review process. Submitted manuscripts, supplementary files, editorial correspondence, author information accessible through OJS, reviewer invitations, and review records are confidential.

This prohibition applies even when an external service describes itself as private unless the tool has been explicitly approved by IJoDS or INSIGHT for use with confidential editorial material under appropriate contractual and technical safeguards.

AI must not replace a reviewer's independent scientific assessment. Reviewers remain fully responsible for evaluating originality, methodological rigor, validity, significance, clarity, and suitability for IJoDS.

Reviewers may use AI for non-confidential supporting tasks, such as improving the language of review text that they themselves have written or conducting a general background literature search, provided that no confidential manuscript content is supplied to the tool.

Material AI assistance in preparing a review report should be disclosed to the editor. Basic spelling, punctuation, or grammar checking does not normally require disclosure.

These requirements supplement the IJoDS Peer Review Processes and Conflicts of Interest / Competing Interests policies.

14. Use of AI by Editors and Editorial Staff

Editors, Editorial Board members, and editorial staff must protect the confidentiality of submitted manuscripts and the integrity of the double-blind review process.

Confidential manuscripts, peer-review reports, unpublished figures or data, author identities, reviewer identities, and non-public OJS records must not be uploaded to external generative-AI tools unless the tool has been explicitly authorized by IJoDS or INSIGHT for that purpose with suitable privacy, confidentiality, data-security, and contractual safeguards.

AI systems must not make autonomous editorial decisions. Decisions concerning desk evaluation, reviewer selection, interpretation of reviewer reports, revision requests, acceptance, rejection, correction, expression of concern, or retraction remain the responsibility of human editors.

Any AI-assisted system authorized for technical or administrative support must function only as decision support. Human editorial judgment and accountability must remain decisive, consistent with the IJoDS Journal Management and Peer Review Processes policies.

15. Documentation and Editorial Verification

Authors should preserve reasonable documentation of material AI use when that documentation is relevant to methodological reproducibility, disclosure, or research integrity.

Depending on the study, records may include:

• tool and model names or identifiers;
• versions and dates of access;
• prompts or prompt templates that materially affected research outputs;
• relevant generated outputs;
• configuration and sampling settings;
• logs or workflow records;
• human validation and correction procedures; and
• evidence linking generated visualizations or analyses to underlying data.

IJoDS may request relevant documentation, underlying data, code, original images, or other research materials when reasonably necessary to assess reproducibility, disclosure compliance, or a research-integrity concern. Such requests will be handled in accordance with the journal's Data and Reproducibility, Allegations of Misconduct, and Privacy policies.

16. Non-Disclosure, Misuse, and Corrective Action

IJoDS distinguishes between responsible, transparently disclosed AI assistance and AI use that compromises scholarly integrity.

Potentially serious breaches include:

• fabrication or falsification of data, evidence, results, references, or citations;
• deceptive manipulation of research images;
• materially misleading or undisclosed AI-generated scientific content;
• misrepresentation of authorship or scholarly contribution;
• breaches of manuscript confidentiality by reviewers, editors, or staff;
• unlawful or unethical disclosure of protected personal or proprietary information;
• copyright or licensing violations; and
• intentional concealment of AI use when disclosure is required to evaluate the integrity or reproducibility of the work.

Depending on the nature, seriousness, and stage of the case, IJoDS may request clarification or documentation, require revision or correction, reject a manuscript, initiate a publication-ethics investigation, publish a correction or expression of concern, retract an article, or take another proportionate action under its established policies.

Suspected misconduct will be handled under Allegations of Misconduct. Post-publication action will follow Post-Publication Discussions and Corrections. Authors who believe an AI-related editorial or procedural decision was handled incorrectly may use the journal's Complaints and Appeals process.

17. Relationship with Other IJoDS Policies

This policy forms part of the IJoDS publication-ethics system and must be read together with the following journal policies. Where requirements overlap, the provision that provides the stronger protection for research integrity, transparency, confidentiality, privacy, and accountability should normally be followed.

18. Policy Review and Reference Basis

AI technologies and scholarly publishing practices continue to develop rapidly. IJoDS will review this policy periodically and may revise it to reflect changes in technology, publication ethics, data-protection expectations, research-integrity standards, and recognized scholarly publishing practice.

This policy is part of the IJoDS Publication Ethics framework and has been developed with reference to current guidance on responsible generative-AI use in scholarly publishing, including:

• Elsevier, Generative AI policies for journals (current policy includes guidance for authors, images and artwork, reviewers, and editors).
• Committee on Publication Ethics (COPE), Core Practices.