Jurnal Pengembangan Penyuluhan Pertanian (JP3) recognizes that artificial intelligence (AI), including generative AI and AI-assisted tools, may support research, manuscript preparation, language editing, data analysis, and scholarly publishing. The journal adopts this policy to promote transparency, accountability, research integrity, and the responsible use of AI throughout the publication process. This policy is informed by the COPE Discussion Document: Artificial Intelligence (AI) in Decision Making and the COPE Core Practices.

AI tools may be used as assistance in scholarly work, but their use must not compromise the accuracy, originality, transparency, confidentiality, or ethical integrity of a manuscript. Authors, reviewers, and editors remain responsible for their own scholarly and editorial decisions. COPE's discussion of AI in publication decision-making highlights the ethical questions associated with AI-supported publishing processes and the importance of considering the role of human oversight. See the COPE Discussion Document on AI in Decision Making.

Authors may use appropriate AI tools to assist with language editing, grammar correction, translation, outlining, coding support, or other legitimate research and writing activities, provided that the use does not violate research ethics, confidentiality, copyright, data protection, or other applicable requirements. Authors must verify all AI-assisted content before submission and remain responsible for the accuracy and integrity of the entire manuscript. This policy is informed by the COPE resources on Authorship and Contributorship.

AI tools must not be listed as authors or co-authors because they cannot assume the responsibilities and accountability required of human authors. Human authors must be identifiable, must approve the submitted manuscript, and must accept responsibility for their contributions. See the COPE Position Statement on Authorship and AI Tools.

Authors must disclose the use of generative AI or other substantive AI-assisted tools when such use contributes to the development of the manuscript, research design, data processing, analysis, interpretation, or presentation of research findings. The disclosure should identify the tool, explain how it was used, and describe the relevant parts of the research or manuscript that were affected. This disclosure requirement is a JP3 editorial policy designed to support transparency and accountability, informed by the COPE resources on Authorship and Contributorship and the COPE Core Practices.

Routine spelling correction, grammar checking, or other basic language assistance that does not materially contribute to the intellectual content of the manuscript may be exempted from disclosure at the journal's discretion. When in doubt, authors should disclose the use of the tool and describe its purpose. Authors must follow any more specific disclosure requirements stated in the current JP3 Author Guidelines.

Authors must independently verify AI-generated statements, factual claims, calculations, statistical interpretations, quotations, and references. AI-generated outputs must not be treated as authoritative evidence without verification against reliable sources. Authors must not fabricate research data, invent references, misrepresent findings, manipulate images, or present AI-generated material as empirical evidence that was not actually obtained. These requirements are consistent with the COPE Core Practices on Data and Reproducibility, Intellectual Property, and Misconduct Allegations.

When AI tools are used in research design, data collection, data processing, statistical analysis, qualitative coding, or interpretation, authors should describe the tool and its role in sufficient detail for readers to understand how the results were produced. Authors must assess the suitability and limitations of the tool, verify outputs using appropriate methods, and ensure that the use of AI does not compromise the validity or reproducibility of the research. These expectations are informed by the COPE Core Practices on Data and Reproducibility.

For agricultural extension and agricultural development research, authors must pay particular attention to the accuracy of AI-assisted interpretations of farmer interviews, survey responses, field observations, agricultural measurements, and community data. Authors must not use AI-generated information as a substitute for actual field observations, valid measurements, or properly documented research evidence.

Authors must disclose AI-generated or materially AI-modified images, figures, illustrations, or other visual materials when their use is relevant to the scholarly content of the manuscript. Such materials must not misrepresent research observations, experimental results, field conditions, or original data. Authors are responsible for ensuring that the use of these materials complies with applicable copyright, licensing, consent, and research integrity requirements. This policy is informed by the COPE Core Practices on Intellectual Property and Data and Reproducibility.

Reviewers must treat submitted manuscripts and related editorial materials as confidential. Reviewers must not upload or disclose confidential manuscripts, reviewer reports, unpublished data, or identifiable author information to external AI tools where doing so could compromise confidentiality or violate the journal's policies. Reviewers who are uncertain whether an AI tool may be used must seek guidance from the editor before using it. These requirements are informed by the COPE Ethical Guidelines for Peer Reviewers and the COPE Discussion Document on AI in Decision Making.

Reviewers remain responsible for the originality, accuracy, fairness, and scholarly quality of their assessments. AI-generated review comments must not replace the reviewer's independent expert judgment, and reviewers must verify any suggestions or factual claims produced by AI before including them in a report. Reviewers must not delegate the confidential assessment of a manuscript to an AI system in a manner that breaches the journal's confidentiality requirements. See the COPE Ethical Guidelines for Peer Reviewers.

Editors may consider appropriate AI-assisted tools for administrative or screening support, such as manuscript classification, formatting checks, language assessment, or identification of potential text overlap, provided that their use is consistent with the journal's policies and applicable confidentiality requirements. AI outputs must be treated as assistance rather than conclusive evidence of misconduct or an automatic basis for an editorial decision. COPE discusses the opportunities and ethical challenges of AI-assisted editorial processes in its Discussion Document: Artificial Intelligence (AI) in Decision Making.

Editors must exercise independent judgment when evaluating manuscripts and must not rely solely on AI-generated recommendations to accept, reject, or otherwise decide the outcome of a submission. Potential misconduct identified by automated or AI-assisted tools must be assessed fairly and in context before any action is taken. These principles are informed by the COPE Discussion Document on AI in Decision Making and the COPE Core Practices on Peer Review Processes and Allegations of Misconduct.

Undisclosed or inappropriate use of AI may be considered an editorial concern when it involves fabricated information, misleading attribution, plagiarism, manipulated data, false references, or a failure to disclose AI use where disclosure is required by the journal. The journal will assess the nature and significance of the concern before deciding on an appropriate response. The handling of suspected misconduct should follow fair and transparent procedures informed by the COPE Core Practices.

If concerns arise regarding AI use, the journal may request an explanation, relevant disclosures, supporting documentation, or clarification from the authors. Where credible concerns about research or publication misconduct remain unresolved, the journal may follow appropriate editorial procedures and, where necessary, refer the matter to the relevant institution. If an issue is identified after publication, the journal will consider appropriate measures to maintain the accuracy of the scholarly record. These procedures are informed by the COPE Core Practices on Allegations of Misconduct and Post-Publication Discussions and Corrections.

By submitting a manuscript to JP3, authors confirm that all statements, data, references, figures, and other materials in the submission have been reviewed and verified by the responsible human authors. Authors must disclose AI use as required by this policy, ensure that AI tools are not credited as authors, and accept responsibility for the content and integrity of the submitted work. These requirements are informed by the COPE Position Statement on Authorship and AI Tools and the COPE resources on Authorship and Contributorship.

JP3 will review this policy periodically to reflect developments in AI technology, research practices, and publication ethics. Authors, reviewers, and editors are responsible for following the current version of the policy published on the journal website. This policy is a journal-specific framework informed by COPE resources; the COPE Discussion Document on AI in Decision Making is a discussion document and should not be interpreted as a formal, universally binding COPE policy.