MAIN MENU

ARTICLE TEMPLATE
VISITOR
Substantia permits responsible use of generative artificial intelligence where it does not replace human scholarly judgment or compromise research integrity. Authors remain fully accountable for every claim, reference, quotation, analysis, translation, illustration, and conclusion. An AI tool cannot be listed as an author and cannot approve a manuscript or assume responsibility for its accuracy.
Authors must disclose substantial assistance from generative AI, stating the tool name and, where available, version; the purpose of use; the manuscript sections or outputs affected; and the steps taken to verify accuracy. Routine spelling, punctuation, or grammar assistance that does not alter intellectual content does not require the same disclosure. When an AI system forms part of the research method, its role, settings, inputs, evaluation, and limitations must be described in the Method section.
AI must not be used to fabricate data, invent interviews or participant statements, generate fictitious citations, simulate findings as real observations, or conceal unattributed borrowing from published works. Synthetic images, translations, textual excerpts, and illustrative examples must not be represented as authentic historical documents, manuscripts, or field evidence.
Quotations from Qur'anic verses, hadith, tafsir works, classical texts, and academic references must be checked against reliable source editions and fully attributed. AI-generated transliteration, translation, bibliographic details, and claims about transmission or textual history require independent verification. The existence of a plausible-looking reference is not evidence that a source exists or supports the claim.
Authors must protect personal data, confidential research materials, and intellectual property when interacting with AI services. Editors and reviewers may not upload unpublished manuscripts or confidential review reports to public generative AI services. Editorial and peer-review judgments must remain independent human assessments. Any suspicion of inappropriate AI use is evaluated on specific evidence, with an opportunity for author explanation; detection scores alone do not establish misconduct.
