Artificial Intelligence Policy

This policy sets out how authors, reviewers, and editors of Journal Geram (Gerakan Aktif Menulis) may use large language models (LLMs), such as ChatGPT, and other generative artificial intelligence tools. It follows the position of the Committee on Publication Ethics (COPE) on AI and authorship, and the disclosure format widely used in scholarly publishing.

Use of LLMs and Generative AI in Writing a Submission

The journal recognizes that LLMs and generative AI can be useful productivity tools when preparing a manuscript. They can help generate initial ideas, structure content, summarize, paraphrase, and refine language. However, language models have inherent limitations and cannot replace human creativity, judgment, and critical thinking. Human oversight and a substantive intellectual contribution remain indispensable to ensure that what readers receive is accurate, sound, and appropriate.

Authors who use LLMs therefore take the following limitations into account:

ObjectivityGenerated text may reproduce biases in earlier published content, including racism, sexism, and other biases, and minority viewpoints may be under-represented. Because the output is decontextualized, these biases are harder to detect.
AccuracyLLMs can produce false content, especially on complex or ambiguous topics. They may generate text that reads well but is wrong, give incorrect facts, misquote literary texts, misattribute sources, or invent citations that do not exist. Authors verify all data, quotations, and citations against the primary sources. Some models also lack recent information.
ContextLLMs struggle with context, idiom, sarcasm, humour, and figurative language, which matter greatly in language and literature research. This can lead to errors and misinterpretation.
Training dataModel performance depends on large amounts of high-quality training data, which may be limited for some fields and for regional languages, reducing the tool’s reliability.

Guidance for authors

Authors clearly state any use of a language model in their manuscript, naming the tool and the purpose. This applies to drafting, translation, data analysis, coding, and image creation.

Disclosure. The statement is placed at the end of the manuscript, before the reference list, in a section entitled “Declaration of AI and AI-Assisted Technologies in the Writing Process”:

During the preparation of this work the author(s) used [NAME OF TOOL / SERVICE] in order to [REASON]. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

The declaration does not apply to basic tools for checking grammar, spelling, or references. If there is nothing to disclose, no statement is needed.

Authors who use these tools also:

Verify the accuracy, validity, and appropriateness of the content and citations generated, and correct any errors or inconsistencies.
Keep a list of the sources used to produce content and citations, including those suggested by a language model, and check every citation.
Watch for plagiarism when a model reproduces substantial text from other sources, and cross-check the original sources. Generated text is still subject to the 20 percent Turnitin limit described in Screening for Plagiarism.
Acknowledge the limitations of language models in the manuscript, including possible bias, errors, and gaps in knowledge.
Do not present AI-generated material as data, quotations, or references that they have not verified.

AI tools such as ChatGPT cannot be listed as authors, because they cannot take responsibility for the work. Appropriate corrective action will be taken if a published article is found to contain undisclosed use of such tools, under the Allegations of Misconduct and Correction, Retraction, and Withdrawal policies.

Editors and Reviewers

The integrity of editorial work and peer review depends on confidentiality, objectivity, and accuracy at every stage. With the growing use of generative AI, clear boundaries are needed.

Reviewers. A manuscript under review is confidential and must not be uploaded to, shared with, or processed by a generative AI tool in any form. Doing so may breach the authors’ confidentiality, infringe copyright, and violate data privacy. Reviewers do not use generative AI to write, structure, or assist in preparing their reports, and remain fully responsible for the content and conclusions of their evaluation.
Editors. Manuscripts and editorial correspondence are confidential. Editors do not upload manuscripts, decision letters, or other editorial communications into generative AI tools, even to polish language. Editorial decisions rest on the editor’s own professional judgment, not on AI output.
Permitted technical use. AI-based technology may be used for limited technical functions, such as similarity checking or format checks, provided it runs within secure, official systems that respect data privacy and ethical standards.
Suspected misuse. If there is reasonable suspicion that an author or reviewer has used generative AI without disclosure, the editor reports it to the Editor-in-Chief for investigation.

The journal supports the responsible use of new technology, but the scholarly assessment of research and editorial decision-making remain human responsibilities that are not delegated to machines.