Article Sidebar
Main Article Content
Abstract
Publishers and journals increasingly require transparency when authors use generative artificial intelligence
(GenAI), but the operational details of disclosure remain uneven. This exploratory document analysis compares
ten purposively selected publisher- or journal-level policy units against twelve criteria. Coding categories were
explicit (E), conditional or partial (C), not explicit in sufficiently reviewed text (N), and unverified (U) when
adequate primary-source evidence was unavailable. In response to reviewer-identified discrepancies, we
rechecked the relevant official policy texts and updated the matrix and evidence register together. The revised
matrix contains 79 E, 24 C, 17 N, and 0 U cells out of 120. The documents suggest recurring attention to human
accountability and limits on AI authorship, while disclosure detail, location, language editing, research-stage use,
images, and peer-review confidentiality vary across the selected documents.