VALIDITAS DATA HADIS DALAM PRAKTIK TAKHRIJ HADIS MENGGUNAKAN CHATGPT

THE VALIDITY OF HADITH DATA IN THE PRACTICE OF TAKHRIJ AL-HADITH USING CHATGPT

Abstract

This study aims to assess the validity of ḥadīth data generated through the practice of takhrij using ChatGPT. The research background stems from the growing use of generative language models as research tools, while the authenticity of the data they produce, especially in the discipline of ḥadīth, has rarely been examined systematically. The research employed a qualitative descriptive method with a content-analysis approach. The findings indicate that ChatGPT can identify source compilations and present Arabic texts rapidly, and in many cases provide an assessment of the ḥadīth’s status (ṣaḥīḥ or ḥasan). However, its accuracy and consistency in handling the chains of transmission (sanad) do not meet the standards of classical takhrij. The language model does not trace all transmission routes, evaluate the reliability of narrators, or detect anomalies (shudhūdh) and hidden defects (ʿillah). These results affirm that ChatGPT is suitable only as an initial tool for literature discovery, while final verification must still be carried out through traditional takhrij methods by qualified scholars. The study recommends the formulation of ethical guidelines and the development of an integrated ḥadīth database combined with artificial intelligence to support responsible teaching and academic research.