Plagiarism and Artificial Intelligence Policy

Plagiarism and Artificial Intelligence Policy

Originality and Similarity Screening

Jurnal Konstruksi dan Infrastruktur (JKI) requires original manuscripts with accurate attribution of text, ideas, data, figures, tables, and other material obtained from existing sources.

Submitted manuscripts undergo similarity screening. An overall similarity level above 25% may result in a request for clarification or revision. This percentage is a screening guideline, not an automatic determination of plagiarism.

Editors examine the sources, distribution, extent, attribution, and significance of overlapping material. A manuscript below the guideline may still contain plagiarism; a higher score may include legitimately quoted or attributed material.

Substantial unattributed copying, misleading paraphrasing, and undisclosed reuse or duplication are unacceptable. Findings may lead to revision, rejection, or post-publication action under the Research Misconduct and Corrections and Retractions policies.

AI Use by Authors

AI tools cannot be authors. Authors remain responsible for the accuracy, originality, references, analysis, and conclusions of their manuscripts.

Use beyond routine spelling, grammar, language correction, or formatting must be disclosed, identifying the tool, purpose, and affected sections. AI used as part of a research method must be described sufficiently for evaluation and reproducibility.

AI must not be used to fabricate or misrepresent data, images, references, or findings.

AI Use by Reviewers and Editors

Reviewers must not use generative AI to analyze or assess manuscripts or produce review comments. Manuscripts or portions of manuscripts must not be entered into AI applications for peer review.

Editors must not use AI to replace scientific assessment or editorial decisions, or to generate editorial assessments. Any permitted administrative or language assistance must protect confidentiality and be disclosed to the Editor-in-Chief with its purpose and scope.

Confidential manuscripts, unpublished data, reviewer reports, and editorial correspondence must not be uploaded to external generative AI services.