ChatGPT as an Algorithmic Epistemic Mediator in Contextual Islamic Religious Education and Religious Moderation

This study critically examines ChatGPT as an algorithmic epistemic mediator in contextual Islamic Religious Education (PAI) learning and its implications for religious moderation. The development of generative artificial intelligence has transformed how students access, interpret, and validate religious knowledge, raising pedagogical and epistemological concerns in Islamic education. This study employed qualitative library research by analyzing scholarly literature published between 2021 and 2026 from academic databases and journal portals, including Google Scholar, SINTA, Scopus-indexed journals, Springer, Taylor & Francis, and Elsevier. The literature was analyzed through reduction, thematic categorization, critical interpretation, and conceptual synthesis. The findings indicate that ChatGPT has an ambivalent role in contextual PAI learning. On the one hand, it may support religious moderation by expanding access to diverse Islamic perspectives, encouraging student inquiry, strengthening digital religious literacy, and facilitating reflective learning. On the other hand, it may also create risks such as simplification of religious teachings, algorithmic bias, pseudo-authority, and pseudo-moderation when AI-generated responses are accepted without critical verification. This study argues that ChatGPT should not be positioned as an autonomous epistemic actor or a final source of religious authority, but as a mediated learning aid that requires teacher verification, critical digital literacy, and reflective-dialogical learning. The study contributes to Islamic education scholarship by offering a conceptual framework for integrating ChatGPT responsibly into contextual PAI learning without weakening the epistemological foundations of religious knowledge.