Illusory Citations and Academic Hallucination: Generative AI and the Erosion of Scholarly Authenticity — A Comparative Study of Arab and International Open-Access Journals

This study aims to analyze the impact of risks associated with the use of generative artificial intelligence tools on scholarly integrity in research published in open-access journals, by conducting a comparative analysis across three levels: the global level, the level of developed countries, and the Arab region. The study adopts a critical analytical methodology based on a systematic literature review (23 previous studies) and the analysis of open-access data from platforms such as OSF Preprints, CrossRef, and Semantic Scholar . The findings reveal that the proportion of entirely fabricated hallucinated citations reaches 19.9%, with significant errors present in 45.4% of other citations (Enago, 2026). At the Arab level, the results indicate a clear knowledge gap: while 84.2% of researchers call for mandatory disclosure of AI use, the percentage of Arab academic libraries that have implemented clear ethical policies does not exceed 12%. The study concludes that open-access journals are facing an unprecedented “crisis of trust,” and proposes an integrated ethical governance framework combining transparency, accountability, institutional training, and automated citation verification.