Artificial Intelligence dalam Transformasi Digital Perbankan Syariah: Systematic Literature Review tentang Kualitas Layanan, Efisiensi Operasional, dan Kepatuhan Syariah

Abstract

Digital transformation has compelled Islamic banks to integrate Artificial Intelligence (AI) as a core component of their service architecture. Grounded in the Technology Acceptance Model (TAM) and Maqashid Syariah, this study examines the role of AI in Islamic banking digital transformation, focusing on customer service quality, operational efficiency, and Sharia compliance. A qualitative Systematic Literature Review (SLR) guided by the PRISMA protocol was employed, analyzing 23 articles selected from 145 initial records across Scopus, DOAJ, and Google Scholar through CASP-based quality appraisal (Cohen's κ = 0.81). Findings reveal that AI enhances service responsiveness, algorithmic personalization, and operational efficiency, yet introduces structural challenges including technostress, digital literacy gaps, trust deficits among high-religiosity segments, and algorithmic opacity that undermines substantive Sharia compliance. Two original frameworks are proposed: the Human-in-the-Loop Sharia model (HILS), a layered oversight architecture integrating AI-driven compliance screening with Sharia Supervisory Board interpretation; and the AI-Integrated Sharia Quality Framework (AISQ), comprising four simultaneous conditions, inclusive responsiveness, equitable efficiency, layered compliance, and trust construction required for AI to function as a legitimate instrument of Islamic banking transformation. These contributions advance the literature beyond descriptive assessments toward a normative-operational synthesis guiding institutional implementation and regulatory design.