Ethical Pluralism in AI Governance: A Comparative Analysis of Maqasid al-Shariah and International Frameworks in Islamic Finance

Abstract

Artificial Intelligence (AI) is being adopted at speed in Muslim-Majority Societies (MMS), above all in digital and Islamic financial services. The international instruments that now shape AI governance (the UNESCO Recommendation on the Ethics of AI, the NIST AI Risk Management Framework and the OECD AI Principles) have, however, very little to say about Islamic ethical and legal norms. In this paper we ask whether and on what conditions, selected objectives of Maqasid al-Shariah can be combined with these frameworks to govern AI in digital entrepreneurship and Islamic finance. The design is qualitative and comparative. We coded a purposive corpus (classical and contemporary Islamic legal sources, international AI governance instruments, Shari'ah and regulatory standards issued by AAOIFI and the IFSB and peer-reviewed studies) against a predefined grid of six ethical themes and five dimensions of comparison and then set the resulting propositions against an illustrative regulatory analysis of Malaysia and Algeria. In practice the two normative sources agree on harm prevention, accountability, transparency and dignity. Where they part ways is on the origin of ethical limits and on whether those limits may be balanced against expected benefits. Building on this, we propose the Maqasid AI Filter Framework (MAFF), a Maqasid-AI Compliance Index (MACI) with certification tiers and a three-lines institutional model for Shari'ah AI auditing and we discuss enforcement and the stakeholders involved. The contribution is conceptual; it has yet to be tested empirically.