REKONSTRUKSI AKUNTABILITAS AKUNTANSI SYARIAH DI ERA ARTIFICIAL INTELLIGENCE: TELAAH LITERATUR TENTANG AMANAH, TRANSPARANSI, DAN MAQASID SYARIAH

Abstract

The application of artificial intelligence (AI) in Islamic accounting not only has implications for efficiency and transparency but also raises fundamental issues concerning trust (amanah) and accountability. From an Islamic perspective, accountability is a moral responsibility inherent in humans as mukallaf (legally accountable beings), and therefore cannot be fully transferred to a non-mukallaf system. This study aims to analyze the implications of AI for the practice of amanah in Islamic accounting, examine the compatibility between algorithmic transparency and the principle of Shariah transparency, and develop maqasid al-shariah as a framework for evaluating AI accountability. The study employs a qualitative approach through a narrative literature review of accredited national scientific journals, reputable international journals, and official reports from relevant institutions. The study analyzed data thematically and conceptually by integrating literature on Islamic accounting, AI ethics, and Islamic legal philosophy. The findings indicate that AI reconfigures the practice of amanah at the instrumental level, while moral responsibility remains with humans. Algorithmic transparency is also not entirely identical to Shariah transparency, as procedural openness does not necessarily guarantee clarity of substance and fairness in decision-making. Furthermore, maqasid al-shariah can serve as an evaluative framework through three dimensions: maslahah-mafsadah (benefit and harm), justice (‘adl), and the transcendental dimension. These findings reveal an accountability gap, namely the discrepancy between the technical capabilities of AI and the moral responsibility that remains inherent in humans. Therefore, the application of AI in Islamic accounting requires strengthening accountability theory, governance that integrates the Shariah Supervisory Board with AI governance mechanisms, and the development of AI standards oriented toward maqasid al-shariah