The Paradigm Conflict and Convergence Between Artificial Intelligence and Evidence Law: A Study Through EU Regulation and Global Case Law

XINYIPENG, YUESHANHUA

Keywords: evidence law; artificial intelligence; algorithmic accountability; transparency; due process

Paper Type: Short papers / work-in-progress

TL;DR: A comparative study shows how EU-style, risk-based rules can modernise evidence law for the AI era.

Abstract:

The integration of artificial intelligence (AI) into judicial fact-finding has triggered a global paradigm crisis in the law of evidence. This article, centred on the regulatory framework of the European Union Artificial Intelligence Act (‘EU AI Act’), critically analyses the dual character of AI in the evidentiary process, as both an enabler and a disruptor, through the lens of landmark cases from the United States, the United Kingdom and Canada.
The article argues that the EU AI Act, through its risk-based classification, transparency obligations, and fundamental rights impact assessments, provides a systemic blueprint for reforming evidentiary doctrine worldwide. Yet, from State v Loomis to Regina v A, leading cases have revealed a profound tension between legal frameworks and technological realities: algorithmic opacity undermines the right of cross-examination, data bias corrodes the fairness of trial, and generative AI destabilises the foundation of authenticity.
Ultimately, the article contends that the future of evidence law lies in developing a model of responsive regulation, which, by drawing on the EU’s regulatory logic, integrates mandatory algorithmic auditing, dynamic evidentiary disclosure, and algorithmic due process into a coherent governance framework. This, it is argued, is how evidence law can regain vitality in the digital age.

Submission Number: 10