Artificial Intelligence Policies

Qaumiyyah: Jurnal Hukum Tata Negara permits responsible, transparent AI assistance while requiring human authorship, independent scholarly judgment, and protection of confidential material. These rules apply to authors, reviewers, editors, and editorial staff. They draw on COPE's position on authorship and AI tools and relevant ICMJE recommendations, adapted to the journal's legal scholarship and publication process.

1. Scope and Human Accountability

This policy covers generative AI, large language models, chatbots, and AI-assisted tools used in research, manuscript preparation, peer review, editing, or production. AI systems cannot be authors, co-authors, reviewers, or decision-making editors. Responsibility remains with the people performing those roles. Disclosure does not make an otherwise unacceptable practice permissible.

2. Requirements for Authors

  • Permitted assistance. Authors may use AI to support language revision, translation, organisation, preliminary literature discovery, coding, or analysis, provided they disclose the relevant use, verify its results, and retain control of the work. The manuscript must represent the authors' own research contribution and reasoned argument.
  • Independent verification. Check every AI-assisted factual statement, quotation, citation, translation, calculation, and interpretation against reliable original sources. An AI response is not a substitute for the source it purports to describe.
  • Legal and Islamic legal sources. Verify constitutional provisions, legislation, judicial decisions, legal doctrines, and Islamic legal materials in the relevant authoritative texts. Check jurisdiction, amendments, dates, article or paragraph numbers, and the context of quotations. Do not rely on invented cases, statutes, fatwas, or references.
  • Prohibited practices. AI must not be used to fabricate or falsify evidence, interview statements, research participants, observations, results, quotations, or references; conceal plagiarism; manipulate authorship; or present machine-produced reasoning as the authors' independent analysis without substantive human evaluation.
  • Rights and data protection. Authors must protect participants, confidential records, third-party intellectual property, and unpublished material entrusted to them. Check an AI service's retention, training, and reuse terms before supplying material, and obtain any required permission. Removing names alone may not remove identifying information.

3. Disclosure and Research Documentation

At submission, describe generative AI assistance in the cover letter or comments to the editor and in an AI Assistance Declaration within the manuscript. Identify the tool, provider, model/version where available, date of use, purpose, affected sections, and the authors' verification. Include updated disclosure when revisions introduce additional AI use. Relevant assistance must remain disclosed in the published article.

Where AI forms part of the research method, also explain its role in the Methods or equivalent section, including relevant prompts, settings, input selection, validation, and limitations. Retain sufficient records for editorial inquiry and reproducibility, subject to confidentiality obligations. Clearly identify simulated or synthetic material when its use is methodologically justified; never present it as observed evidence.

Suggested declaration: Tool and model/version: [details]. Provider and date: [details]. Purpose and affected sections: [details]. Verification and revisions performed by the authors: [details]. All authors approve the final manuscript and accept responsibility for its content.

Routine non-generative spellchecking, formatting, and reference management do not require an AI declaration. Generative rewriting, translation, substantive content assistance, and AI-supported research methods do. AI-assisted conceptual diagrams must be identified in their captions and checked by the authors; fabricated documentary or empirical images are prohibited.

4. Requirements for Reviewers

Reviewers must conduct their own assessment of originality, method, evidence, interpretation, and contribution. They must not delegate the review or its recommendation to AI. AI assistance does not replace either of the journal's minimum two independent human reviewers or its double-blind review process.

Obtain prior written permission from the handling editor before using generative AI for any review-related assistance. Any permission is limited to a clearly defined supportive task involving non-confidential material, such as refining generic wording. Disclose the tool, purpose, and extent of assistance to the editor, and personally verify the report. Concerns about an author's AI use must be explained to the editor with specific evidence, not an unsupported allegation.

5. Requirements for Editors and Editorial Staff

Editors remain responsible for screening, reviewer selection, interpretation of reports, and all editorial decisions. AI must not autonomously accept, reject, rank, or determine the scholarly merit of a submission. Editors must evaluate disclosures in context and must not penalise authors merely for transparent, permitted assistance.

Editorial AI assistance must be authorised by the editor-in-chief, confined to a documented supportive purpose, and checked by a human editor. The journal must inform authors of any substantive AI assistance introduced during copyediting or production; authors must be able to review resulting changes. Suspected misuse must receive individual assessment and an opportunity for explanation.

6. Confidentiality and Review Integrity

Reviewers, editors, and staff must not upload, paste, or transmit submissions, extracts, figures, datasets, review reports, editorial correspondence, or identifying information to public or external generative AI services. This restriction applies even when names have been removed. Approval for limited non-confidential assistance does not authorise disclosure of manuscript material. No AI use may expose identities, appropriate unpublished ideas, or compromise double-blind review.

7. Assessment, Breaches, and Corrections

Compliance is assessed through the purpose and extent of assistance, disclosure, source verification, confidentiality, and human responsibility. The journal does not treat an AI-detection percentage as conclusive proof of misconduct or apply it as an automatic rejection threshold. Editors may request clarification, underlying records, or revisions and must consider the response.

Depending on the evidence and seriousness, an established breach may lead to revision, rejection, reassessment of a review, or correction of the published record. Retraction, institutional referral, or other action must follow the journal's applicable ethics procedures and relevant COPE guidance. Allegations and appeals should be addressed through the journal's official contact information.

Guidance Informing This Policy

Policy adopted: 3 November 2026. This policy will be reviewed as publishing guidance and AI technologies develop.