Generative AI Policy

Generative artificial intelligence (AI) tools, such as ChatGPT, Gemini, Copilot, and Claude, can support scholarly work, but they also carry risks. This policy sets out how authors, editors, and reviewers of Efisiensi may use them. It follows the COPE position statement on authorship and AI tools and the 2026 national journal accreditation guidelines, and will be revised as practice develops.

In brief. AI may assist, but people remain fully accountable. AI cannot be an author. Authors and editors must disclose any use. Reviewers must not use AI in reviewing. Unpublished manuscripts must never be uploaded to AI tools.

1. Why Caution Is Needed

  • Inaccuracy and bias: AI can produce false statements, fabricated references, and biased content.
  • Lack of attribution: AI output may reproduce others' ideas or text without credit.
  • Confidentiality: text entered into AI platforms may be stored or reused by the provider.
  • Research data: survey responses, interview transcripts, and organisational documents may contain personal or sensitive information.

2. Policy for Authors

2.1 Accountability and authorship

  • Authors are fully responsible for the originality, validity, and integrity of the whole manuscript, including any part produced with AI assistance. They must check all AI output and every reference against the original source.
  • AI tools must not be listed or cited as authors. They cannot take responsibility for the work, consent to publication, or manage copyright.

2.2 Permitted uses (with disclosure)

  • Language editing, proofreading, and translation of the authors' own text.
  • Brainstorming and outlining.
  • Searching for and organising literature, provided every source is verified.
  • Help with code or formulas for analysis (for example R, Python, or spreadsheet formulas), provided the authors check and validate the results.

2.3 Prohibited uses

  • Submitting AI-generated text as the authors' own analysis or argument without substantial human revision.
  • Generating or altering data, results, interview transcripts, quotations, or survey responses.
  • Generating references. Every reference must be a real source that the authors have read.
  • Creating or manipulating images presented as data, such as photographs, documents, or screenshots.
  • Uploading identifiable participant data to AI tools without consent and ethics approval for that use.

2.4 Disclosure

Every manuscript includes a Declaration of Generative AI Use, stating the tool and version, how and why it was used, and how the authors checked the output. If no AI was used, the authors state: “The authors did not use generative AI tools in the preparation of this manuscript.” Spelling checkers and reference managers do not need to be declared.

Example. “During the preparation of this manuscript, the authors used ChatGPT (GPT-4o, OpenAI) to improve the language of the Introduction and Discussion. The authors reviewed and edited all output and take full responsibility for the content. No AI tool was used to generate data, analyse data, or produce references.”

3. Policy for Reviewers

  • Reviewers must not use generative AI to analyse, assess, summarise, or write comments on a manuscript.
  • Reviewers must not upload a manuscript under review, or any part of it, to an AI tool. This would breach confidentiality and the anonymity of double-blind review.
  • Reviewers remain fully responsible for their reviews.
  • Reviewers who suspect undisclosed AI use in a manuscript should tell the editor in their confidential comments.

4. Policy for Editors

  • Editors must not upload unpublished manuscripts, reviews, or decision letters to generative AI tools.
  • Editors must not use AI to make or justify editorial decisions. Decisions are made by people.
  • Editors may use tools approved by the journal for specific tasks, such as similarity checking or reference validation, and must disclose such use when asked.
  • AI-detection results are not proof of AI use. Authors are given a chance to explain before any action is taken.

5. Handling Breaches

  1. Identification: a concern is raised by an editor, reviewer, reader, or screening (for example, non-existent references).
  2. Explanation: the authors are asked to explain and provide evidence, such as drafts, data, or sources, within 14 days (manuscripts under review) or 30 days (published articles).
  3. Action: undisclosed but acceptable use requires a declaration (before publication) or a correction (after publication). Prohibited use leads to rejection or retraction, and may be handled under the Research Misconduct policy. Reviewers or editors who breach this policy are removed from the reviewer database or editorial team.

Sources

  • Committee on Publication Ethics. (2023). Authorship and AI tools: COPE position statement. https://publicationethics.org/cope-position-statements/ai-author
  • Direktorat Jenderal Sains dan Teknologi. (2026). Keputusan Direktur Jenderal Sains dan Teknologi Nomor 374/DST/D.D1/HM.01.01/2026 tentang Petunjuk Teknis Akreditasi Jurnal Ilmiah. Kementerian Pendidikan Tinggi, Sains, dan Teknologi.
  • International Committee of Medical Journal Editors. (2024). Recommendations for the conduct, reporting, editing, and publication of scholarly work. https://www.icmje.org/recommendations/