AI as Consistency Aid, Not Decision Maker: A Post-Divorce Alimony Decision Model in Religious Courts

Abstract

Debates on artificial intelligence (AI) in the judiciary remain at the level of macro-regulation, leaving concrete problems in Religious Court courtrooms unaddressed. One such problem is the determination of post-divorce maintenance, namely iddah, mut’ah, past maintenance (nafkah madhiyah), and child support related to hadhanah, which rests on judicial discretion and is therefore prone to yield different amounts in similar cases. This study aims to formulate a model that positions AI not as a decision-maker but as an aid to the consistency of judicial reasoning. It uses a normative-juridical and conceptual approach, examining Supreme Court Regulation (PERMA) No. 3 of 2017, Supreme Court circular letters on post-divorce maintenance, Islamic legal maxims, and maqasid al-shari’ah, and tests the design against evidence on human–AI interaction in courts. The analysis yields a three-step model: (1) AI presents relevant considerations and ranges of amounts, with source decisions, drawn from previous decisions; (2) the judge determines the amount through ijtihad with written reasons, explaining any departure from the range; and (3) the reasoning of the decision includes a statement on the use of AI. The model’s novelty lies in separating information from decision, in a comply-or-explain rule that seeks consistency of reasoning rather than uniformity of amounts, and in a testable evaluation protocol. The study concludes that the model can safeguard judicial independence and ijtihad, strengthen legal certainty and justice for women and children, and be applied under existing law, provided that verified data, safeguards against anchoring and over-reliance, and institutional oversight are in place.