Abstract
The integration of Artificial Intelligence (AI) into Islamic education presents both opportunities and significant challenges, particularly with the emergence of the usul.ai platform as an AI agent for research into classical texts (turāts). This study aims to map the technical characteristics and epistemological functions of usul.ai, as well as to explore the learning experiences of first-year students (cohort 2025) at the Faculty of Islamic Studies (FDI) at UIN Syarif Hidayatullah Jakarta in integrating usul.ai into their studies. Furthermore, this study identifies fundamental limitations and pedagogical risks associated with the use of AI within the Islamic higher education ecosystem. This research employs a qualitative method with a phenomenological design to understand the participants’ direct experiences. Data were collected through Focus Group Discussions (FGDs) with first-year students who had interacted with usul.ai, and were subsequently analysed using thematic analysis and data triangulation to ensure the validity of the findings. The research findings indicate that usul.ai acts as a catalyst that accelerates classical text literacy and research efficiency amongst students. The platform functions as adaptive scaffolding that bridges language barriers, though it still requires close supervision. Theoretically, this study proposes a hybrid model of ‘shuhbatu ustadz’ (teacher guidance) and ‘shuhbatu AI’ (AI support), in which AI is positioned as Khadim al-‘Ilm (servant of knowledge). Although beneficial, the use of AI carries pedagogical risks in the form of data hallucinations and academic dependence, which may erode critical thinking skills. Therefore, this study recommends that the integration of AI into Islamic education must remain ‘human-in-the-loop’ to preserve the authenticity of the sanad and the values of tarbiyah.