Generative AI Policy

Jurnal Ilmu Manajemen recognises that generative AI can support scholarship, and also that it introduces risks to accuracy, attribution, and confidentiality. The guiding principle is simple: AI may assist the scholar, but it cannot replace the scholar. This policy follows the COPE position statement on authorship and AI tools, the ICMJE recommendations, and WAME guidance, and will be revised as practice develops.

Risks the policy addresses

• Fabrication. Generative tools invent facts, statistics, and references that look plausible and are not real.
• Attribution. AI output may reproduce the work of others without credit, which can amount to plagiarism.
• Confidentiality. Text entered into a third-party tool may be stored, reviewed by the provider, or used to train a model.
• Bias. Model outputs can carry bias that matters in research on gender, ethnicity, religion, and vulnerable groups.

Policy for authors

Accountability. Authors are fully responsible for the entire manuscript, including any part produced with AI assistance. They must verify every fact, figure, and reference against the primary source, ensure the argument is their own, and check that the tool's terms protect any confidential or participant data.

Authorship. An AI tool cannot be an author or co-author and cannot be cited as one.

Permitted, with disclosure: brainstorming and outlining; language editing, including support for authors whose first language is not English; translation of the author's own text, carefully checked; literature searching and classification, with every source verified at the original; help with code, provided the code is tested.

Not permitted: submitting unedited AI text as the author's own analysis; generating or altering data, interview transcripts, quotations, survey responses, or field notes; creating synthetic data without a transparent and declared method; generating references; producing abstracts or keywords that are not verified; uploading identifiable participant data to any AI tool without consent and ethics approval.

Images and figures. AI assistance is acceptable for data visualisations, conceptual illustrations, and flow diagrams such as PRISMA, provided it is declared and the underlying data are accurate. It is not acceptable for photographs, documentary evidence, or screenshots, and any manipulation of such images is prohibited.

Declaration. Every manuscript includes a Use of Generative AI statement in the end matter giving the tool and version, the sections affected, the purpose, and how the output was verified. Example:

The authors used ChatGPT (GPT-4o, OpenAI) to improve the readability of the Introduction and Discussion. All AI-assisted text was reviewed, edited, and verified by the authors, who take full responsibility for the content of the article.

Where no generative AI was used, state: "No generative AI tools were used in the preparation of this manuscript." Spelling and grammar checkers and reference managers do not require declaration.

Policy for reviewers

Reviewers must not upload a manuscript, or any part of it including figures and supplementary files, to any generative AI tool, since this breaches confidentiality. They must not use AI to analyse, summarise, or write the review report. AI may be used only to improve the language of a review the reviewer has already written, with no manuscript content entered. Reviewers remain fully responsible for their reports. A suspicion of undeclared AI use should be raised confidentially with the editor, not asserted in the review.

Policy for editors

Editors keep submissions and review materials confidential and never upload them to AI tools. AI is not used to make or justify editorial decisions. Only journal-approved tools may be used for defined tasks such as similarity checking, reference validation, and image-integrity screening. AI-detection output is treated as an indicator, never as proof, and authors are always given the opportunity to explain.

Handling breaches

1. Identification. A concern may arise from screening, a reviewer, an editor, or a reader after publication.
2. Explanation. Authors are asked to explain their use of AI and provide evidence within 14 days for a manuscript under review, or 30 days for a published article.
3. Outcome. Undeclared but otherwise acceptable use is resolved by adding a declaration before publication, or by a corrigendum afterwards. Prohibited use leads to rejection or retraction and may be pursued as research misconduct, including notification of the authors' institution. A breach by a reviewer or editor results in the review being disregarded and the person being removed from the reviewer database or the editorial board.