COMPUTATIONAL LEGAL DIAGNOSTICS AND THE LIMITS OF HUMAN LEGAL REASONING

Abstract

Recent scholarship on artificial intelligence and law has primarily examined the use of AI to support legal decision-making, automate legal tasks, or expand access to legal services. This article advances a different analytical orientation by conceptualizing AI as a diagnostic instrument for examining the internal structure of legal systems themselves. Drawing on interdisciplinary insights from law, socio-legal studies, and computational analysis, the article develops the concept of computational legal diagnostics to describe the use of AI techniques in identifying normative inconsistencies, regulatory overlaps, and enforcement gaps that are difficult to detect through doctrinal legal reasoning alone. Rather than treating law as a set of rules to be applied, this approach analyzes law as a complex institutional system characterized by layered norm production and uneven implementation. Through illustrative examples drawn from a multi-tiered legal context in the Global South, the article demonstrates how AI-assisted analysis can reveal structural patterns of legal fragmentation and institutional bias embedded in contemporary regulatory frameworks. The article concludes by discussing the implications of this diagnostic use of AI for legal reform, institutional design, and the governance of AI in law, emphasizing that such systems should complement rather than replace human legal interpretation and democratic accountability.