The Four Doctrines of Provenance Law - Restoring Evidentiary Integrity and Legal Accountability in the Age of AI.

This paper is a revised edition of the version first published on 25 November 2025. Version 2.0 was revised for academic review on 26 March 2026. Generative AI has introduced a new provenance problem for law. Where text is drafted, revised, interpolated, or semantically shaped by AI, legal systems may face greater difficulty in treating that text as a reliable proxy for human intention, authorship, and accountability. This paper develops four doctrines to analyse that problem: AI Contamination, Human Authorship and Responsibility, Evidentiary Integrity, and Algorithmic Discrimination. It argues that current AI governance frameworks, while important, do not by themselves resolve the upstream legal question of whether AI-mediated text can still be attributed to a human source with sufficient confidence for evidentiary and legal purposes. The paper therefore proposes provenance assurance as an emerging form of legal infrastructure for preserving evidentiary integrity, procedural fairness, and accountable decision-making in the age of AI.