JOVISHE: Journal of Visionary Sharia Economy, published by Yayasan Lembaga Studi Makwa (Makwa Foundation), recognises that artificial intelligence (AI) can support scholarly inquiry, research communication, and technological innovation. At the same time, the use of AI-assisted technologies raises important questions concerning originality, the reliability of evidence, attribution, fairness, confidentiality, data protection, and accountability. This policy establishes clear obligations for all participants in the publication process.

The policy applies to authors, corresponding authors, editors, Editorial Board members, peer reviewers, journal managers, copyeditors, production personnel, and the publisher. It complements JOVISHE's Authorship Criteria, Author Guidelines, Peer Review Process, Reviewers' Guidelines, Publication Ethics, Plagiarism Policy, Conflict of Interest Policy, and Research Misconduct Policy. Its principles are informed by the COPE position on AI and authorship, the January 2026 ICMJE Recommendations, and Indonesia's 2026 scientific-journal accreditation technical guidelines.

For this policy, AI-assisted technologies include generative AI, large language models, image and media generators, automated writing systems, and AI-enabled applications capable of creating or materially transforming scholarly text, code, figures, data, or analysis. The relevant ethical consideration is the function performed, not merely the commercial name of the tool.

Routine tools that correct spelling, punctuation, or minor grammar without adding substantive intellectual content are distinguished from generative tools that develop arguments, synthesise literature, rewrite substantial passages, create analysis, or generate research materials. Material or substantive AI use must be declared. A study that investigates AI itself, or legitimately employs machine-learning methods, is governed by the normal standards of research design, documentation, reproducibility, and ethical oversight in addition to any applicable disclosure requirements.

Authorship is reserved for identifiable human contributors who satisfy JOVISHE's Authorship Criteria. An AI model, chatbot, or automated agent must not appear as an author or co-author because it cannot approve the final manuscript, accept responsibility for the research, declare competing interests, or enter into publication agreements.

• Research ownership. Authors must independently formulate and justify their research questions, disciplinary argument, interpretation, and conclusions. AI-generated output is not a substitute for scholarly reasoning.
• Verification. Every author remains responsible for the truthfulness, originality, statistical accuracy, methodological soundness, citations, and publication rights of all submitted content, including AI-assisted material.
• Source integrity. Authors must verify every quotation, citation, DOI, regulatory provision, data value, and reference against a traceable primary or reputable scholarly source. A plausible-looking AI-generated reference is not evidence that the source exists.
• Collective approval. The corresponding author must ensure that all co-authors are informed of substantive AI use and approve the final declaration and manuscript.

The journal does not impose a blanket ban on responsible AI assistance for authors. The acceptability of a particular use depends on transparency, human oversight, the rights governing the source materials, the reproducibility of affected methods, and the absence of misleading scholarly claims.

Authors must report substantive AI use at the initial submission stage and update that information whenever relevant use occurs during revision. The corresponding author must disclose the use of AI in the submission information or cover letter and in the manuscript under the heading Declaration of AI and AI-Assisted Technologies, ordinarily placed among the end-of-article declarations before the References section.

A sufficiently informative declaration identifies the tool or service, version where available, purpose, tasks undertaken, affected manuscript components, and human verification. Where AI is integral to data collection, processing, quantitative modelling, qualitative coding, image analysis, or research design, its use must also be explained in the Methods section with enough detail to support critical appraisal and appropriate reproducibility.

For consistency and clarity, JOVISHE also asks authors who made no substantive use of generative AI to include a brief negative declaration. Routine spelling and punctuation correction alone does not require a detailed tool-by-tool disclosure; nevertheless, the authors remain responsible for the final language and meaning.

JOVISHE publishes research spanning Islamic economics, Islamic banking and finance, Islamic social finance, ethical enterprise, digital entrepreneurship, halal industries, and sustainable development. AI can be a legitimate research instrument or object of inquiry in these fields, provided its application is scientifically justified and methodologically transparent.

Studies using AI-based forecasting, natural-language processing, sentiment analysis, credit-risk modelling, fintech analytics, e-commerce recommendations, or analysis of zakat and waqf administration must distinguish genuine research observations from simulations or generated data. Authors should describe relevant datasets, inclusion criteria, preprocessing, model selection, performance evaluation, sources of bias, and limitations. Where appropriate, they should also consider fairness, financial privacy, institutional permission, and Islamic ethical implications.

• Empirical studies: Clearly identify measured data, generated features, data splits, evaluation metrics, and the basis for conclusions.
• Qualitative research: Disclose any AI-assisted coding or transcription that materially affects analysis; protect participant anonymity and verify quotations against original recordings or transcripts.
• Systematic and bibliometric reviews: Report the search, screening, verification, and synthesis procedures; AI-derived summaries must not replace assessment of retrieved source literature.
• Design-based or applied innovation: Describe AI-enabled artefacts, evaluation protocols, user or organisational outcomes, and evidence of benefits or limitations rather than merely presenting application screenshots.

Authors must not invent Qur'anic references, hadith citations, scholarly attributions, Islamic legal opinions, regulations, DOI identifiers, institutional statistics, or bibliography entries through AI assistance. If AI is the object of methodological investigation, identify the tool as an instrument or material as appropriate, rather than crediting it as a scholarly author.

Authors must comply with research-participant consent, ethical approval conditions where applicable, personal-data protection, institutional research agreements, copyright, database licences, and confidentiality obligations. Identifiable information concerning research participants, banking customers, zakat beneficiaries (mustahik), business clients, or confidential organisational records must not be entered into an external AI service without an adequate lawful basis, permissions, and appropriate safeguards.

The journal expects authors to understand the data-handling conditions of the AI services they choose. Use of a commercial AI system does not relieve authors of their duties to preserve confidential information or secure rights in copyrighted text, data, code, illustrations, and images. Claims of compliance must be supported by the actual conditions of use rather than assumptions about a product.

Editors and editorial personnel must protect the confidentiality of unpublished manuscripts and the independence of scholarly decisions. Generative AI must not act as an editor, determine novelty or validity, generate substantive peer-review judgments, rank authors on unsupported criteria, or decide whether a paper should be accepted, revised, or rejected.

• Human decision-making: Screening, reviewer appointment, assessment of reviewer reports, research-integrity determinations, and final editorial decisions are made by qualified human editors.
• Confidentiality: Unpublished manuscripts, reviewer reports, protected author identities, correspondence, and confidential editorial files must not be uploaded to public or unapproved AI tools. Exceptional authorised technical processing requires appropriate confidentiality safeguards and, where necessary, explicit permission from affected authors.
• Limited operational support: The editorial office may use authorised technical services for administrative assistance or similarity and integrity screening when appropriate. Such outputs are preliminary indicators, not final judgments, and must be assessed by human staff.
• Editorial and production transparency: Any substantive AI use in editorial or production workflows must be recorded as appropriate and disclosed to authors where relevant. Routine non-confidential administrative work should not create misleading claims about manuscript evaluation.
• Publisher independence: No AI-assisted editorial decision may be influenced by author fees, sponsor interests, commercial arrangements, or conflicts of interest. Responsibility cannot be transferred to a tool provider.

During a review assignment, reviewers must not:

• Use generative or AI-assisted systems to read, summarise, translate for assessment, analyse, judge, or evaluate a submitted manuscript, its data, figures, methods, or conclusions.
• Use AI tools to draft, generate, paraphrase, revise, or complete peer-review comments, numerical scores, methodological critiques, recommendations, or confidential notes to the editor.
• Copy, paste, photograph, upload, forward, or otherwise disclose any part of the manuscript, supplementary materials, reviewer report, or confidential editorial communication to an AI system.
• Delegate a confidential review or any part of the scholarly evaluation to an AI agent, automated service, or unauthorised third party.

A reviewer who cannot comply with these requirements must notify the handling editor and decline or withdraw from the assignment. The reviewer remains personally responsible for the originality, scholarly quality, confidentiality, and timeliness of any review submitted. Any suspected breach will be considered under the journal's Publication Ethics and Research Misconduct procedures.

JOVISHE may request clarification or supporting documentation when it encounters unreliable references, unexplained inconsistencies, misleading generated figures, undeclared use of AI, or concerns about the integrity of the underlying data. Editorial inquiries will focus on verifiable evidence and the material effect of the conduct, not on superficial writing patterns.

AI-detection percentages, automated similarity outputs, and classifier-based estimates of machine-generated prose are not conclusive evidence of misconduct and will not, on their own, justify rejection, disciplinary action, or a public allegation. Editors must examine the specific concern, offer the author an appropriate opportunity to respond, and document a reasoned determination.

1. Initial assessment: The handling editor identifies and records the relevant concern and checks whether there is credible supporting evidence.
2. Author clarification: The corresponding author may be asked for an accurate AI disclosure, original sources, methodological explanations, provenance records, or corrected material relevant to the concern.
3. Independent evaluation: A non-conflicted editor assesses the explanations and, where appropriate, seeks independent scholarly or institutional advice without breaching confidentiality.
4. Proportionate action: Outcomes may include clarification, revision, correction of a disclosure, further methodological assessment, rejection, or referral under a formal misconduct procedure where warranted.
5. Procedural record: Questions, responses, evidence, decisions, and dates must be maintained accurately in the journal's authorised editorial records.

If a material AI-related concern is identified after publication, JOVISHE will evaluate whether the scholarly record requires a clarification, correction, expression of concern, or retraction. The appropriate action depends on the reliability of the work and the available evidence; it is not determined automatically by the mere presence of AI assistance or an omitted minor disclosure. Where necessary, the journal may refer potential misconduct to a relevant institution for further investigation.

This policy becomes operative upon formal approval and publication by the JOVISHE editorial office. It applies prospectively to relevant manuscript submissions and editorial activities. Any action concerning earlier publications must be handled transparently under the journal's established correction and research-integrity procedures. The journal may revise this policy in response to changes in recognised publication-ethics guidance, AI technology, and applicable accreditation requirements.

Submission-stage author checklist

• A complete statement of substantive AI assistance is provided at submission, or a clear no-substantive-use statement is included.
• Any AI model integral to the research methodology is reported with appropriate methodological detail.
• All co-authors have approved the disclosure and checked the manuscript's evidence, references, figures, code, and conclusions.
• Personal data and confidential organisational information have been handled lawfully and securely.
• The author declaration is included in the manuscript and updated if AI use changes during revision.

Related JOVISHE policies and submission

Policy reference framework