AI Tool Policy

ARTIFICIAL INTELLIGENCE USAGE POLICY

JURNAL ILMU KOMPUTER DAN AGRI-INFORMATIKA (JIKA)

A. General Provisions

  1. Jurnal Ilmu Komputer dan Agri-Informatika (JIKA) is committed to upholding academic integrity, originality, transparency, objectivity, and publication quality in computer science, information systems, artificial intelligence, software engineering, and the application of information technology to agriculture, forestry, animal science, fisheries, marine science, food, and the environment.
  2. JIKA recognises that artificial intelligence (AI) technologies, including generative AI, can support literature searching, data processing and analysis, programming, translation, and language editing.
  3. The use of AI is permitted provided it is responsible, transparent, verifiable, and does not replace the intellectual contribution, scientific reasoning, and academic accountability of humans.
  4. This policy applies to authors, editors, peer reviewers, and all parties involved in JIKA's publication process, for all manuscript types (research articles, review articles, and short communications).
  5. This policy governs the use of AI as a tool in research and publication. Research that develops, trains, or studies AI as a scientific subject remains welcome and is subject to the specific requirements in Section 9.
  6. This policy is based on the guidance of the Committee on Publication Ethics (COPE), IPB Academic Senate Regulation No. 47 of 2023, and the Guidelines for the Use of Artificial Intelligence (AI) in the Implementation of the Tridharma at IPB University (2025).

B. Definitions and Examples of AI Tools

For the purposes of this policy, the following terms have the meanings set out below.

Term Meaning
Artificial intelligence (AI) Systems, platforms, or software that use techniques such as machine learning, natural language processing, and deep learning to generate, analyse, translate, summarise, or modify text, numbers, code, images, or audio.
Generative AI AI that produces new content, such as large language models (LLMs) and image-generation models.
Basic language tools Spelling, grammar, and punctuation checkers that do not generate new scientific substance.
Substantive use Use of AI that affects the content of a manuscript: drafting or rewriting text, or generating code, images, tables, analyses, interpretations, literature summaries, or translations.
Synthetic data Data generated by an algorithm or model rather than obtained from real measurements, observations, or respondents.
Third-party AI service A cloud-based AI service operated by an external provider that may store or process user inputs under the provider's own terms.

AI tools covered by this policy include, but are not limited to:

  • Text-based generative AI: ChatGPT, Gemini, Claude, Copilot, DeepSeek, and other LLMs.
  • Coding assistants: GitHub Copilot, Cursor, Codeium, and similar tools.
  • Writing, paraphrasing, and translation tools: Grammarly, DeepL, Quillbot, and similar tools.
  • AI-powered literature search and summarisation tools: Elicit, Consensus, Scite, SciSpace, and similar tools.
  • Image generation or editing models: DALL·E, Midjourney, Stable Diffusion, and similar tools.
  • Data analysis, statistics, and visualisation software that uses AI.

C. Core Principles

Every provision of this policy derives from COPE's six Core Principles.

COPE principle Application to AI use at JIKA
Accountability Humans, not AI, are responsible for all manuscript content, assessments, and editorial decisions.
Fairness, justice, and equity Suspected AI use is handled fairly and is never judged on writing style or author origin alone.
Integrity AI must not be used to fabricate data, references, or results, or to disguise plagiarism.
Rigour and accuracy All AI output is verified against data, code, and primary sources before inclusion.
Stewardship The scholarly record is protected, including through correction or retraction when AI misuse is found.
Transparency AI use is clearly disclosed so that editors, reviewers, and readers can evaluate it.

D. Permitted Uses of AI

Authors may use AI as a supporting tool, provided every output is checked and verified before it is included in the manuscript. Permitted uses include:

  1. Improving grammar, spelling, punctuation, sentence structure, and readability.
  2. Assisting with translation and scientific language editing, including from Indonesian into English, with verification by the authors.
  3. Helping to outline or structure the article.
  4. Helping to format references and citations according to JIKA's house style.
  5. Supporting literature searching, mapping, and organisation (e.g., bibliometric analysis), with every source verified.
  6. Assisting with writing, debugging, documenting, and refactoring code, with testing by the authors.
  7. Assisting with processing, cleaning, clustering, and visualising legitimate research data.
  8. Supporting, but not replacing, statistical analysis, modelling, and preliminary interpretation of results under the researchers' supervision.
  9. Exploring initial ideas, problem formulations, and methodological alternatives, which the authors then examine critically.
  10. Creating simple illustrations or conceptual diagrams, disclosed in accordance with Sections 8 and 11.

AI does not relieve authors of their obligation to construct the scientific argument, analyse the findings, and draw conclusions from accountable evidence.

E. Prohibited Uses of AI

The following uses of AI constitute breaches of publication ethics and are handled under Section 14:

  1. Generating all or a substantial part of a manuscript with AI and submitting it as the authors' own work without genuine intellectual contribution and critical review.
  2. Fabricating, altering, or manipulating data, experimental results, model performance metrics, or code outputs that were never actually obtained.
  3. Creating fictitious survey responses, fake interview transcripts, or field data (e.g., crop yields, land imagery, or sensor readings) that were never collected.
  4. Presenting synthetic data as empirical data.
  5. Generating fictitious or unverifiable references, quotations, or DOIs.
  6. Using AI to summarise or paraphrase others' work so as to evade similarity checks or disguise plagiarism.
  7. Translating a manuscript with AI without checking and editing by the authors.
  8. Creating or altering figures, charts, screenshots, or research images with AI in a misleading way or without disclosure.
  9. Generating content that infringes copyright or licences, including code or images owned by others, without permission.
  10. Entering confidential or personal data into third-party AI services without authorisation and adequate safeguards.
  11. Listing AI as an author or co-author, or including it in author contribution statements.

F. Responsibilities of Authors

Authors are fully responsible for all submitted content, including any part generated or assisted by AI. Authors must:

  1. Verify the accuracy, originality, and reliability of all AI-assisted content.
  2. Ensure the manuscript is free from plagiarism, bias, factual errors, and AI hallucinations (information that appears convincing but is false).
  3. Understand, test, and be able to explain any AI-assisted code, analysis, and results.
  4. Attribute the source of any data, quotation, or third-party material used or suggested by AI.
  5. Critically review and edit all AI output so that it meets the scholarly standards of computer science and agri-informatics.
  6. Accept full responsibility for any errors, omissions, or ethical breaches arising from the use of AI.

G. Authorship

In line with COPE's position statement on Authorship and AI tools, AI cannot be listed as an author of an article in JIKA.

  1. AI does not meet the requirements for authorship because it cannot take responsibility for the submitted work. As a non-legal entity, it cannot assert the presence or absence of conflicts of interest, nor manage copyright and licence agreements.
  2. Authorship is granted only to individuals who have made a substantial contribution to the design, conduct, analysis, or writing of the research, have approved the final version, and can be held accountable for its content.
  3. Listing AI among the authors, in author notes, or in author contribution statements is prohibited and may result in desk rejection or retraction.
  4. The corresponding author ensures that all co-authors are aware of and approve the AI use statement in the manuscript.

H. Disclosure of AI Use

Every substantive use of AI must be disclosed. The use of basic language tools does not need to be declared.

H.1 Minimum content of the statement

  1. The name of the AI tool, its developer or provider, and its version or date of access.
  2. The purpose and extent of its use.
  3. The parts of the manuscript or stages of the research that were assisted by AI.
  4. How the authors verified, edited, and intellectually contributed to the AI output.
  5. Confirmation that the authors have reviewed and take responsibility for all AI-assisted content.

H.2 Location of disclosure in the manuscript

Type of use Where to disclose
AI used in collecting, processing, or analysing data; or in generating figures, code, or research models The Methods section, in enough detail for the procedure to be evaluated and reproduced
AI used in drafting, editing, or translating text A dedicated Declaration of AI Use section before the References
All manuscripts The submission form in JIKA's publishing system (mandatory, including when no AI was used)

H.3 Template statements

When AI was used:

Declaration of AI Use. During the preparation of this article, the authors used [tool name, developer, version/date of access] to [purpose and extent, e.g., edit the English of the Introduction and assist in writing data pre-processing scripts]. All AI outputs were reviewed, tested, and edited by the authors against the research data and analysis. The authors take full responsibility for the originality, accuracy, and integrity of the entire content of this article.

When no generative AI was used:

The authors declare that no generative artificial intelligence technology was used in conducting this research or preparing this article.

I. AI as a Research Subject or Method

Manuscripts that develop, train, evaluate, or apply AI models must meet the following transparency and reproducibility standards.

  1. Model description: architecture, pre-trained base model and its version, hyperparameters, and training or fine-tuning procedure.
  2. Dataset: source, licence, size, labelling process, training/validation/test split, and measures taken to prevent data leakage.
  3. Evaluation: metrics, comparison baselines, number of runs or random seeds, and measures of uncertainty (e.g., standard deviation or confidence intervals).
  4. Computation: hardware (CPU/GPU), training time, and software libraries with their versions.
  5. LLMs as a method: when an LLM is used for data annotation, classification, information extraction, or evaluation, report the model name, version or date of access, the full prompts, parameters (e.g., temperature), and validation against human annotation.
  6. Code and data availability: authors are encouraged to share code, prompts, and configurations through an open repository (e.g., GitHub, Zenodo, or OSF). Where these cannot be shared, the reason must be given in the Data Availability Statement.
  7. Synthetic data: augmented or model-generated data (e.g., GAN or diffusion-model images of plants) must be clearly declared with their proportion and reported separately from empirical data.
  8. Limitations and bias: authors discuss model limitations, potential bias (e.g., class or regional imbalance), and the risks of field deployment, particularly for farmers and agricultural stakeholders.

J. Research Data and Privacy

Data processed with AI must be legitimate, verifiable, and protected in accordance with applicable law, including Indonesian Law No. 27 of 2022 on Personal Data Protection.

  1. Data must come from legitimate sources and match the research method. Authors must be ready to provide data and supporting documents when requested by the editor, subject to confidentiality.
  2. Personal data (identities of respondents, farmers, students, or employees) must be anonymised before being processed with AI.
  3. Confidential or strategic data (internal company or agency data, land ownership records, precise locations of research plots, unreleased genetic or variety data) must not be entered into third-party AI services without the data owner's permission and adequate safeguards.
  4. For sensitive data, authors are encouraged to use locally run models or institutional infrastructure.
  5. Where data come from human subjects, informed consent covers processing with AI tools, and research ethics approval is stated in the manuscript.
  6. Spatial and image data (satellite, drone, or IoT sensor) must state their source, licence, and acquisition date.

K. Literature, Citations, Images, and Code

K.1 Literature and citations

  1. AI may be used to identify topics, keywords, and literature maps, but every reference must exist, be relevant, and accurately reflect the original source.
  2. Authors must read and verify every source, including DOI, title, authors, year, and pages.
  3. AI is not treated as a scholarly source. Theories, concepts, and prior findings must be cited to traceable primary sources.

K.2 Images and visualisations

  1. Illustrative images created with AI (e.g., conceptual diagrams) must be labelled in the figure caption, naming the tool used.
  2. Images that constitute research data (experimental images, maps, system screenshots, plant images) must not be created or altered with generative AI, unless this is part of the method and explained in the Methods section.
  3. Reasonable technical adjustments (brightness, contrast, cropping) are permitted if applied uniformly to the whole image and they do not alter the information.

K.3 Program code

  1. Code written or assisted by AI must be understood, tested, and executable by the authors, who are responsible for its correctness and security.
  2. Authors must ensure that AI-generated code does not infringe third-party software licences.

L. Originality and Plagiarism Prevention

Manuscripts must be original. AI assistance with paraphrasing or writing does not remove the obligation to cite others' ideas and work correctly.

  1. Authors must not use AI to disguise plagiarism, circumvent similarity checks, or present others' ideas as their own.
  2. Manuscripts must meet JIKA's similarity index threshold of no more than [__]%.
  3. Similarity check results are one input to the evaluation, not the sole basis for determining plagiarism.
  4. JIKA does not rely solely on AI-detection software to conclude that a breach has occurred, because its accuracy is limited and it can disadvantage authors writing in a second language. Every assessment involves human judgement, other evidence (e.g., fictitious references or data inconsistencies), and direct communication with the authors.

M. Use of AI by Editors, Reviewers, and the Editorial Team

Editors and reviewers remain fully responsible for their assessments and decisions. AI must not replace their scholarly judgement.

  1. Unpublished manuscripts, data, code, review reports, and editorial correspondence are confidential and must not be uploaded to third-party AI services that may store or reuse those inputs.
  2. Reviewers must not draft review reports with generative AI. Language polishing of a self-written report is permitted, provided no manuscript content is entered into an AI service.
  3. Editors and reviewers assess AI use statements as part of the ethical and methodological evaluation of the manuscript.
  4. Editors may use tools approved by the journal management for administrative tasks, such as similarity checks, reference formatting, or initial completeness screening.
  5. JIKA does not use AI to make autonomous editorial or peer-review decisions. Decisions to accept, revise, or reject are always made by humans.
  6. Substantive use of AI in the review process requires the editor's approval and must be disclosed to the journal management.
  7. If the editorial team adopts AI tools in its workflow in future, this will be announced openly on the journal website and will always involve human verification.
  8. Editors who encounter a review report suspected of being AI-generated will seek clarification from the reviewer and, if necessary, request a replacement review.

N. Handling Suspected Breaches and Consequences

Suspected breaches are handled objectively, proportionately, and in line with COPE flowcharts. Authors are always given the opportunity to respond before a decision is made.

Depending on the severity of the breach, editorial actions may include:

  1. A request for clarification of AI use.
  2. A request to revise the manuscript or add an AI use statement.
  3. A request for raw data, code, prompts, or supporting documents.
  4. Rejection of the manuscript at any stage if a breach of publication ethics is proven.
  5. Publication of a correction or an expression of concern for a published article.
  6. Retraction of the article when a serious breach affects the validity of the research findings.
  7. Notification of the authors' institution or a request for a formal investigation where serious misconduct is suspected.
  8. A ban on future submissions to JIKA for a specified period where misuse is severe or repeated.

Honest disclosure of permitted AI use does not affect the assessment of a manuscript's substance.

O. Appeals and Dispute Resolution

  1. Authors who disagree with an editorial decision related to AI use may submit a formal written appeal to the Editor-in-Chief.
  2. The appeal must state the grounds for disagreement, include supporting evidence (e.g., manuscript version history, code, or data), and refer to the relevant provisions of this policy.
  3. Appeals are reviewed by an internal ethics panel not involved in the original decision and, where necessary, with reference to COPE guidance or the COPE Forum.
  4. The decision on the appeal is communicated to the authors in writing and is final.

P. Policy Review and Closing Provisions

  1. By submitting a manuscript to JIKA, authors confirm that they have read, understood, and agreed to this policy.
  2. Authors who are unsure whether their use of AI complies with this policy are encouraged to contact the editorial office before submission.
  3. This policy is reviewed at least once a year, or sooner in response to changes in technology, COPE guidance, or IPB and government regulations.
  4. Matters not covered by this policy are referred to the JIKA Editorial Board, with reference to COPE guidance and applicable IPB regulations.
  5. JIKA supports the ethical use of AI to improve the efficiency and quality of research in computer science and agri-informatics, without diminishing the human role in advancing knowledge and making scientific decisions.

References