2025/12/18 by Tsybuliak, Natalia, Suchikova, Yana · 1 voice
Medicine · Social Sciences · #Artificial Intelligence in Healthcare and Education #Ethics and Social Impacts of AI #Academic integrity and plagiarism
paper · doi:10.5281/zenodo.17972708
openalex publication_date 2025/12/18 · openalex created_date 2025/12/19 · openalex updated_date 2026/07/01
This presentation was delivered at the 3rd International Conference “Integrity, Open Science and Artificial Intelligence in Academia and Beyond: Meeting at the Crossroads” (IOSAI-2025), held online on December 17–18, 2025. The contribution addresses the growing phenomenon of stigmatization of generative artificial intelligence (GenAI) use in academic research and writing. Despite the widespread and often informal adoption of AI tools for tasks such as translation, editing, structuring, and information synthesis, many researchers remain reluctant to disclose their use due to fears of reputational damage, rejection in peer review, or accusations of compromised academic integrity. The presentation introduces the concept of the “purity myth” in academic writing—the belief that scholarly texts must remain untouched by AI to retain their legitimacy—and demonstrates how this assumption is historically unfounded and counterproductive. Drawing on recent empirical studies, publisher policies, and survey data, the authors show that stigmatization drives AI use underground, penalizes transparency, reinforces bias in evaluation, and ultimately undermines research quality and reproducibility. As a constructive alternative, the presentation advocates a shift from binary questions (“Was AI used?”) toward delegation-based transparency (“What tasks were delegated to AI?”). In this context, the GAIDeT (Generative AI Delegation Taxonomy) is presented as a practical and scalable framework for responsible AI disclosure across the full research lifecycle, enabling clarity, accountability, and comparability without conflating AI assistance with authorship. The contribution is situated at the intersection of academic integrity, open science, and research governance, and is particularly relevant for researchers, editors, reviewers, research managers, and policy-makers seeking evidence-based approaches to integrating generative AI into scholarly practice while preserving trust in science.