<![CDATA[Hadith studies, as the second pillar of Islamic jurisprudence, are fundamentally anchored in the strict principles of authenticity and precision regarding sanad (chain of narration) and matan (text) data. While generative Artificial Intelligence (AI) offers efficiency, it introduces a critical methodological vulnerability known as "hallucination"-the tendency to generate fictitious yet convincing data. This research aims to analyze how AI hallucination specifically manifests as a methodological threat and data distortion in Hadith studies. Employing a descriptive analysis method with a comparative case study using a Large Language Model (LLM), the study compared AI responses concerning Hadith data (Sahih al-Bukhari from the Ibn Kathir print) and its commentary (syarah) (Ibn Daqiq Al-'Ied) against their primary sources. The findings identified three categories of distortion: Contextual Hallucination (misinterpreting "Dar Ibn Kathir Publisher" as "Imam Ibn Kathir"); Factual Hallucination (providing incorrect matan text for a specific Hadith number, No. 6309); and Linguistic Hallucination (fabricating academic-sounding but fictitious syarah analysis, such as the false claim of using the word (دعائم) and inventing a descriptive analysis by Ibn Daqiq). This research concludes that AI hallucination poses a real and verifiable threat to Hadith data integrity, urging urgent mitigation strategies, including improving training data quality and involving humans such as hadith experts to validate the data provided by AI.]]>
Copyrights © 2026
