<![CDATA[The rapid adoption of conversational artificial intelligence for religious guidance has created an unprecedented situation: Growing numbers of Muslims now consult AI systems for guidance on Islamic legal questions, effectively treating algorithms as muftis. This paper examines why LLM-based AI systems cannot fulfill this role through three interconnected analyses: examination of the digital Islamic knowledge ecosystem showing systematic visibility bias toward unqualified sources, articulation of an eight-domain framework distinguishing what muftis possess from what AI lacks, and empirical testing of six AI platforms on tayammum (dry ablution) across the four Sunni madhāhib. All systems - including those designed specifically for Islamic content - produced errors reflecting fundamental mismatches between AI architectural objectives (linguistic plausibility, content visibility, user satisfaction) and what Islamic legal authority requires (factual verification, scholarly authentication through sanad, epistemological humility). While specialized systems demonstrate that accuracy is technically achievable through a combination of agentic capabilities with curated data, general-purpose systems accessible to most users lack these safeguards and will systematically fail at representing the authenticated transmission and hierarchical legal reasoning structures madhāhib developed over fourteen centuries.]]>
Copyrights © 2026
