AI, retractions, and research integrity: An emerging challenge for libraries

Abstract

The rapid integration of artificial intelligence (AI) tools into scholarly research and publishing has introduced new challenges to research integrity, manifesting in a sharp rise in AI-related retractions. This conceptual review examines the emerging typologies of unethical AI use in academic publishing, including AI-generated textual content with hallucinated citations, AI-assisted data fabrication, AI-generated or manipulated images and figures, AI-enabled peer-review manipulation, and failures to disclose AI involvement. With 764 AI-related retractions identified and a marked spike in 2023, these developments signal systemic vulnerabilities in scholarly communication systems. The paper explores the critical role of librarians as stewards of trustworthy knowledge, educators in AI literacy, and partners in research integrity governance. Librarians are uniquely positioned to detect AI-related misconduct through metadata expertise, to curate tools that flag problematic content, and to guide researchers on ethical AI use aligned with publisher policies and the Committee on Publication Ethics (COPE) guidelines. The review proposes coordinated policy interventions across publishers, journals, libraries, and institutions, emphasizing mandatory AI disclosure, integration of retraction metadata into discovery systems, and capacity-building programs for information professionals. As AI reshapes the research landscape, proactive engagement by librarians is essential to uphold the core values of accuracy, transparency, and accountability in scholarly communication.