Beyond Efficiency: A Procedural and Substantive JusticeFramework for Artificial Intelligence in Indonesia's Judiciary

Abstract

This article extends the author's earlier mapping of the risks and opportunities of judicial artificial intelligence (AI) in Indonesia by asking a narrower, theory-driven question: which conception of justice should govern the normative limits of AI-assisted adjudication, a procedural one or a substantive one? Employing normative legal research with a conceptual-theoretical approach, the study places John Rawls' distinction between pure, perfect, and imperfect procedural justice alongside Tom Tyler's empirical procedural-justice criteria and substantive-justice concerns drawn from Dworkin's law-as-integrity and distributive-justice literature. It argues that judicial AI in Indonesia can, at best, function as an instrument of imperfect procedural justice: no algorithm supplies an independent, verifiable criterion of a substantively correct outcome, so procedural safeguards, explainability, participation, neutrality, and correctability, cannot by themselves guarantee substantive fairness, and must be paired with distinct substantive safeguards addressing bias, equality, and Indonesia's legally pluralistic context of civil law, adat, and Islamic law. The article proposes a dual-track governance model that treats procedural and substantive justice as separate but interlocking pillars, rather than folding both into a single generic regulatory checklist. The findings suggest that Indonesian policy on judicial AI should be organized explicitly around these two tracks to avoid conflating the legitimacy of a decision-making process with the fairness of its outcome.