Large language models for automated Islamic ontology construction and knowledge integration

Introduction: This study addresses the challenge of integrating foundational Islamic resources, particularly the Qur’an and Hadith, into a unified semantic framework. Its primary objective is to automatically unify these resources into a coherent and comprehensive Islamic ontology, enabling a more precise representation of Islamic knowledge and supporting broader semantic applications through advanced artificial intelligence (AI) and ontology engineering techniques. Materials and methods: GPT-4 was applied to the Leeds and King Saud University (LK) Hadith corpus to extract Hadith topics from the six canonical collections. These outputs were systematically organised into a bilingual Arabic–English Hadith_Teaching_Topics (HTT) dataset and subsequently integrated into the development of a web ontology language (OWL)-based ontology, designed to assimilate diverse Islamic resources within a single knowledge representation framework. Results: The proposed approach produced a structured bilingual topic dataset and an automatically generated ontology that offers both machine-readable and human-readable formats. The resulting framework facilitates interoperability, semantic querying, reasoning, and expert validation across integrated Islamic resources. Conclusions: The findings demonstrate the potential of large language models (LLMs) to support automated ontology integration within a specialised religious domain, providing a scalable and semantically robust framework for linking and organising Islamic knowledge.