ALGORITHMIC JUSTICE FROM RULE OF LAW TO RULE OF CODE The El-Rakhawi Doctrine of Computational Adjudication From Human Jurisprudence to Self-Evolving Legal Systems - A Unified Framework for the Age of Autonomous Adjudication, Algorithmic Bias as Crime, and AGI Accountability

Abstract

ALGORITHMIC JUSTICE by Dr. M.K.A. El-Rakhawi (DOI: 10.5281/zenodo.23499006) establishes a unified framework for computational adjudication across seven parts. It transitions from Rule of Law to Rule of Code, recognizing smart contracts as self-executing law and DAOs as algorithmic legal persons. Algorithmic bias is formalized mathematically via disparate impact metrics P(O=1|G=1)/P(O=1|G=0)≠1, with criminal liability requiring mens rea (recklessness/gross negligence) while civil penalties apply strict liability. AGI accountability extends the Agentic Personhood Doctrine with Autonomy(A)=1-(human_control/total_control), employing joint-and-several liability externally and normalized internal attribution for recourse. The Living Law evolves via reinforcement learning π*=argmax_π E[Σγ^t R(s_t,a_t)], where the reward function R=U(justice,efficiency,equity,liberty,security) uses Multi-Attribute Utility Theory (MAUT) to avoid incommensurability paradoxes. Hybrid adjudication distinguishes continuous variables (linear weighting) from categorical decisions (Bayesian veto-power). Constitutional constraints and global governance frameworks ensure democratic legitimacy. Legal protection excludes abstract ideas per Alice Corp. v. CLS Bank (2014), covering only concrete legal-engineering implementations.