E-Prosiding Persidangan Serantau Manasik Haji 2025 (MANASIK2025)

MANASIK2025 Putrajaya, Malaysia, 7-8 Oktober 2025 641 2. Retrieval-First Approach: AI responses cite original texts verbatim before generating summaries. 3. Authority Tags: Each response signals binding status (e.g., “This ruling is binding in Selangor as gazetted on [date]”). 4. Escalation Protocols: Uncertain queries are flagged for referral to human muftis. 5. Audit Trails: All AI interactions are logged for institutional review. 6. Multilingual Interface: Answers available in Malay, English, and Arabic for pilgrims. 5.2 Proposed Model (Textual Description of Figure 1) A diagram (Figure 1) could depict the flow: • User Query → AI System o Step 1: Retrieve from curated corpus (fatwa, resolutions). o Step 2: Apply authority filter (state vs. national). o Step 3: Generate answer with citation + authority tag. o Step 4: If low confidence → escalate to human mufti. o Step 5: Log answer for audit. This model emphasizes that AI is supportive, not substitutive , ensuring alignment with both uṣūl al - fiqh and Malaysian governance. 6. Synthesis of Findings The research objectives are addressed as follows: • Objective 1: The study clarified Sharīʿah and Malaysian legal conditions for valid fatwa, showing that AI must remain jurisdiction-aware and authority-bound. • Objective 2: The analysis identified recurrent Hajj issues (ritual errors, health, women’s rulings, travel constraints) suitable for AI assistance. • Objective 3: The opportunities (timely, multilingual guidance) and risks (hallucinations, erosion of accountability) of AI were systematically evaluated. • Objective 4: A maqāṣid -based conceptual framework was developed, combining curated corpora, retrieval-augmented generation, and human oversight. Collectively, these findings demonstrate that AI-driven fatwa in Malaysia is feasible only under strict governance , with maqāṣid al - sharīʿah as ethical compass and state fatwa authority as institutional anchor. CONCLUSION This study set out to conceptualize a Sharīʿah -compliant framework for AI-driven fatwa automation in the Malaysian Hajj and ʿUmrah context. Through a doctrinal and qualitative analysis, several key findings emerged. First, fatwa issuance in Malaysia is governed by both classical uṣūl al -fiqh principles and state-specific legal structures. Fatwa is valid only when issued by accredited scholars, supported by evidentiary reasoning ( istidlāl ), and gazetted by the ruler. The study highlighted the layered hierarchy of authority in Malaysia, where state-gazetted fatwa are binding, while national Muzakarah resolutions and Hajj- specific guidelines remain advisory unless formally adopted.