Generative AI and AI-Assisted Technologies Policy

Preventif: Jurnal Kesehatan Masyarakat permits responsible use of artificial intelligence to assist scholarly work. Human authors, reviewers, and editors remain accountable for their contributions and decisions.

This policy applies to all manuscript types and associated materials. Institutional, ethical, consent, and data-protection requirements must also be respected.

Human Responsibility and Authorship

AI tools cannot be listed as authors or co-authors. Authors must independently check AI-assisted content for accuracy, bias, originality, and appropriate attribution.

References suggested by AI must be checked against the original publications. AI-generated answers cannot substitute for scientific sources supporting a claim.

Assistance in Manuscript Preparation

Authors may use AI to assist with language, translation, organization, or other preparation tasks, provided that they review the output and disclose relevant use.

AI assistance must not replace the authors’ understanding of the methods, interpretation of findings, or substantive intellectual contribution.

Fabricated participants, observations, results, references, or ethics approvals are unacceptable, whether generated manually or through AI.

AI in Research Methods

When AI is part of the research workflow, describe its role in Methods. Examples include automated literature screening, data extraction, classification, prediction, image analysis, and code development.

Report information sufficient to evaluate the workflow, including:

• Tool or model name and version, where available.
• The task performed and relevant settings.
• Input sources and preprocessing.
• Human checks and validation procedures.
• Important limitations and their potential effects.

For evidence syntheses, explain how AI-assisted screening or extraction was checked and how disagreements or errors were resolved.

Synthetic or simulated data must be explicitly identified and justified. They must never be represented as observations collected from actual participants or experiments.

Disclosure Requirements

Include a section titled “Declaration of the Use of AI” before References and disclose relevant use in the cover letter.

The declaration must identify the tool, purpose, extent of assistance, and human verification. Research-related use must additionally be documented in Methods.

Basic spelling checks that do not generate or substantially rewrite content need not be declared. Generative rewriting, translation, summarization, coding, analysis, and image creation used in the submitted work must be disclosed.

Where no relevant use occurred, include a statement confirming this. Authors should ask the editorial office when uncertain about disclosure.

Figures and Illustrations

AI-generated conceptual illustrations may be considered when they are clearly distinguishable from empirical evidence. Authors must verify scientific content, labels, and numerical information and disclose how the illustration was produced.

AI must not invent or deceptively alter data-bearing images. If AI processing is a legitimate research method, retain the original material and explain and validate the processing.

Scientific diagrams, tables, and graphs must accurately represent the underlying evidence.

Privacy and Confidential Information

Authors must not provide identifiable participant information, restricted datasets, or confidential third-party material to AI services without appropriate authorization and safeguards.

Before using a service, consider its data retention, training, access, and sharing arrangements. Removing names alone may not adequately protect sensitive information.

Reviewers and Editors

Reviewers and editors must protect unpublished manuscripts, associated files, review reports, identities, and correspondence.

Public or third-party generative AI services must not receive these confidential materials. Any proposed use of an institutionally secured system requires prior journal authorization, an assessment of confidentiality safeguards, and disclosure of its use.

AI must not replace independent scientific assessment or determine acceptance and rejection. Reviewers and editors remain responsible for checking their reports and decisions.

Declaration Examples

Adapt the following examples to describe actual practice. Delete inapplicable wording.

Manuscript preparation:

“The authors used [tool and version, where available] to assist with [specific task]. They checked the assisted material against the underlying evidence, corrected errors, and approved the final text. Responsibility for the publication remains with the authors.”

Research workflow:

“[Tool/model] assisted with [research task]. Methods describes the inputs, workflow, validation, and human checks. The authors take responsibility for the resulting analyses and interpretations.”

No relevant AI use:

“No generative AI was used to prepare or substantially revise this manuscript or its illustrations. Any AI-based research methods, if applicable, are described separately in Methods.”

Editorial Assessment and Concerns

Editors may request additional information about an AI-assisted workflow. Undisclosed use or suspected misuse will be assessed according to its significance and the journal’s ethics procedures.

Authors will have an opportunity to explain. Possible outcomes include clarification, revision, rejection, or a post-publication notice where justified.

AI-detection scores must not be treated as conclusive proof of AI use or misconduct. Decisions require contextual assessment and supporting evidence.