Ernest Lim (National University of Singapore — Faculty of Law) and Ilya Akdemir (National University of Singapore — Faculty of Law) have posted Same Words, Different Worlds: The Illusion of Shared Judicial AI Principles on SSRN. Here is the abstract:
Global AI governance frameworks increasingly deploy shared terminology such as “transparency”, “bias” and “fairness”, creating an illusion of alignment across jurisdictions. Through comparative analysis of AI deployment in common law courts (the U.S., the U.K., Australia) and Chinese courts, we expose how seemingly universal concepts mask fundamentally divergent interpretations and implementation. This article makes three distinct contributions to legal scholarship. First, it provides a critical and comprehensive doctrinal and jurisprudential analysis of why and how judicial AI principles operate differently across legal systems. We demonstrate that common law courts define “reliability” as case-specific accuracy verified through adversarial contestation and ex post sanctions, while Chinese courts prioritize systemic reliability via centralized oversight. We show that “bias” threatens individual rights in adversarial systems but endangers social harmony in China. We explain that “transparency” enables party contestation in common law jurisdictions but renders processes legible to supervisory authorities in China. Second, we move beyond comparative law by identifying three extra-legal forces driving these divergences: competing efficiency logics prioritizing different judicial functions; AI as “inherently political technology” presupposing particular institutional arrangements; and archival-epistemic structures determining authority over legal knowledge. We address the paradox of why common law judges deploy AI in sentencing and bail decisions but resist AI deployment for core adjudicatory functions. We also analyze the puzzle as to why they perceive AI as threatening their authority in contrast to Chinese judges. This analysis has broad applicability beyond AI to other transformative technologies in law. Finally, our analysis has important practical implications for global AI governance. We demonstrate why international frameworks cannot succeed through harmonized terminology or technical standards alone. Due to discrepant frames underpinning common law and Chinese perspectives on AI, we suggest that successful global AI instruments require a frame-reflective reciprocal translational approach.
Highly recommended!
To receive a daily summary of posts from Legal Theory Blog by email, get a free subscription to Legal Theory Stack.