Journal Article

Meaningful machine confrontation

Benjamin Welton

2024

Overview

Machine-generated evidence is now ubiquitous in criminal trials, and more sophisticated forms of inculpatory evidence are on the way. Courts have almost universally held that the Confrontation Clause does not give criminal defendants a constitutional right to confront machine-generated evidence, except in narrow cases where the evidence also contains testimonial statements made by a human operator. Several scholars have countered that the Confrontation Clause should be read more broadly to consider machine accusers as \"witnesses\" that trigger confrontation rights. While cross-examination has been the traditional confrontation right in the American legal system, there are machine analogs such as source code disclosure, broadened discovery, machine access, and expanded live testimony that could be afforded to criminal defendants facing machine-generated evidence. This Note presumes that the right to machine confrontation should exist and focuses on which of the proposed alternatives constitutes meaningful machine confrontation in light of the technology at issue. Existing calls for machine confrontation propose blanket solutions without fully considering the practical concerns judges invoke, the realities defense attorneys face, or the types of technology involved. This Note aims to fill this gap with three contributions. First, it provides a comprehensive view of the technologies entering criminal trials, including those sophisticated tools that will one day dominate discussion. Second, it proposes a taxonomy of the technological characteristics underlying these tools, focusing on characteristics that implicate the Confrontation Clause's concerns about dignity and reliability. Third, it considers how these underlying characteristics dictate which proposed confrontation rights would be most meaningful. A machine-specific approach like the one proposed here serves three important functions. First, by considering how machine confrontation would work, it rebuts the judicial argument that such confrontation is impossible or impractical. Second, it offers practical guidance for defendants and cash-strapped attorneys who must face these machines. Third, it sets the stage for other interventions-if judges continue to reject the Confrontation Clause as the vehicle to respond to the problems of machine-generated evidence, legislatures and rulemaking bodies will need guidance on how to structure alternatives.

• Publisher: Stanford University Law School,Stanford Law School,Stanford University, Stanford Law School
• Subject: Artificial intelligence / Attorneys / Confrontation (Criminal law) / Confrontation (Law) / Discovery (Law) / DNA evidence / Evidence / Evidence (Law) / Evidence, Criminal (International law) / Evidence, Expert / Expert evidence / Government regulation / Judges & magistrates / Judicial power / Laws, regulations and rules / Machine learning / Methods / Software / Taxonomy / Technological innovations / Trials / Witnesses