<![CDATA[Anyone with a mobile phone can now ask generative artificial intelligence for religious answers in fluent Indonesian, complete with verses and hadiths, at any hour. This article examines what that change means for religious authority and for adab, the ethics of conduct through which knowledge is transmitted and received in Islamic tradition, by conceptualising such a system as an algorithmic ulama. The study is a qualitative interpretive analysis. Thematic analysis was applied to 24 scholarly works, and document analysis to 15 dated public primary documents, comprising statements of Nahdlatul Ulama, the Indonesian Ulama Council, and Muhammadiyah, together with state guidance, international and interfaith frameworks, and a Malaysian comparator, supported by data on internet use. Three findings emerge. First, generative artificial intelligence reproduces the form of a scholarly answer while disrupting five dimensions of the adab of knowledge: sanad, the verifiable chain of transmission; ṣuḥbah, the formative bond between teacher and learner; tawaqquf, the humility to suspend judgement; ikhtilāf, legitimate scholarly disagreement; and accountability. Second, the major Islamic organisations agree that artificial intelligence may not act as a mufti or a teacher of religion, but they diverge on its use in da'wah and learning, from prohibition, to setting boundaries, to ethical enablement. Third, these responses address users and scholars rather than the systems themselves, so system design remains largely ungoverned. The article therefore proposes an adab-by-design framework that translates the five dimensions into requirements for developers, religious institutions, and users, and argues that adab can give artificial intelligence ethics a concrete vocabulary beyond general principles of transparency and human oversight.]]>
Copyrights © 2026
