The rapid integration of Generative Artificial Intelligence (GenAI) into digital da'wah content engines, including AI-animated short videos and automated religious-consultation chatbots, has changed how Islamic communication circulates in Indonesia, introducing risks such as contextual reductionism, hallucinated citations, and a weakening of the traditional scholarly chain (sanad). Existing scholarship has mostly described institutional adoption, such as Nahdlatul Ulama's Kitab AI and Muhammadiyah's Chat HPT, without proposing a testable, shariah-grounded auditing procedure for the GenAI systems that now sit between a text and its audience. This study addresses that gap by formulating a Maqasid al-Shariah-Based Algorithmic Auditing Model (MAAM) through qualitative document analysis and cyber-hermeneutic reading of publicly available GenAI da'wah content, institutional AI policy documents, and the wider Indonesian and international scholarship on Islamic AI ethics. The audit matrix is anchored in Hifz ad-Din (preservation of religion) and Hifz al-'Aql (preservation of intellect), operationalised through the Qur'anic principles of tabayyun (verification) and qaulan sadidan (accurate speech). The analysis shows current GenAI da'wah engines are optimised for engagement rather than context, producing source stripping, flattened ikhtilaf (interpretive plurality), and occasional fabricated attribution. The proposed model responds with a tri-tier workflow: pre-prompt source authentication, in-pipeline hermeneutic constraints in the system prompt, and post-output human clearance by a certified da'i or editorial board, paired with an institutional certification mechanism. The contribution is conceptual: MAAM is a proposed model, illustrated with a single worked application, whose evidentiary base is documentary and literature-grounded, and whose inter-rater reliability on real content pipelines awaits empirical validation.
Copyrights © 2026