Algoritma di Balik Akad: Rekonstruksi Maqashid Syariah Dalam Kecerdasan Buatan Perbankan Syariah Digital

Abstract

The adoption of artificial intelligence (AI) in digital Sharia banking services including automated credit scoring, robo-advisory, and algorithm-based Sharia compliance detection systems is expanding rapidly; however, an ethical framework grounded in the perspective of Maqashid Sharia (the objectives of Sharia) has yet to be fully formulated. This study aims to critically examine the alignment of algorithmic decision-making architectures in Sharia banking with the five dimensions of Maqashid (hifz al-din, hifz al-nafs, hifz al-'aql, hifz al-nasl, and hifz al-mal) and to formulate an ethical governance framework to address this normative gap. The research employs a descriptive-analytical qualitative approach utilizing a library research design and policy document analysis, complemented by scenario-based case simulations derived from representative industry practices. Primary data comprising fatwa documents, Financial Services Authority (OJK) regulations, and Sharia bank governance reports were analyzed using content analysis and Maqashid based analysis. The findings indicate that generic AI models adopted by Sharia banks tend to replicate conventional data biases oriented toward profit maximization (short-term maslahah dzanniyah), thereby potentially overlooking the dimensions of distributive justice that lie at the core of hifz al-mal and hifz al-nasl. This study proposes a five-layer algorithmic governance model comprising Sharia by design, periodic Maqashid audits, explainable AI transparency, customer objection mechanisms, and the involvement of the Sharia Supervisory Board in the model development cycle as an original contribution to bridging the gap between technological innovation and Sharia objectives within the contemporary Islamic finance industry.