The Blurred Thresholds of AI-Use Disclosure: Health Professions Education Journal Editors' Expectations of Necessity and Sufficiency.

Introduction: Generative AI is a powerful resource for health professions education (HPE) researchers publishing their work. However, questions remain about its use and guidance about disclosure is inconsistent. This study explores journal editors' experiences and expectations of AI-use disclosure, to assist journals to clarify expectations and authors to satisfy them. Methods: In this descriptive qualitative study, editors were interviewed between January 6, 2025, and May 7, 2025 using Zoom. Eligible participants were identified through journal webpages and snowball sampling. A purposive sampling strategy prioritized HPE journals and included a limited sample of general medical journals to explore transferability. Data collection and thematic analysis proceeded iteratively. Results: Eighteen participants, including 9 chief editors and 9 associate/deputy editors were interviewed. Fourteen worked in HPE journals, four in general medical journals. The analysis revealed 4 themes: 1) the basics of disclosure, made up of content expectations and process knowledge; 2) the necessity threshold, regarding which circumstances require disclosure; 3) the sufficiency threshold, regarding how much detail to include; and 4) the factors blurring these thresholds, which included the speed of change, the co-construction of standards, and the uneasy fit of some scientific principles with the AI-use context. Conclusions: While editors shared basic disclosure expectations, these were complicated by blurred thresholds of sufficiency and necessity that may exacerbate uncertainty in the scholarly community. By attending to these thresholds and the factors blurring them, and by working to articulate shared disclosure standards, HPE journals can help authors safely navigate the shifting norms of AI-use disclosure.