the JURY and MACHINE: Artificial Intelligence Limits of Algorithmic Adjudication

This book examines whether artificial intelligence can meaningfully substitute for the jury as the principal fact-finder in the American legal system. Rather than treating the question primarily as a constitutional issue governed by the Sixth and Seventh Amendments, the book approaches it as a problem of epistemology and institutional design: what cognitive and normative tasks does the jury actually perform when it renders a verdict, and can those tasks be replicated by computational systems? Drawing on scholarship in legal epistemology, cognitive psychology, and artificial intelligence, the book argues that the jury is not merely an evidence-weighing mechanism but an epistemic institution that integrates credibility assessments, narrative reasoning, and community moral judgment into authoritative determinations of legally sufficient truth under conditions of uncertainty. Jury verdicts emerge from a collective process of human interpretation structured by adversarial procedure, evidentiary rules, and deliberation-features that reflect the legal system's recognition that fact-finding in trials is fundamentally interpretive and normative rather than purely computational. The book further argues that although contemporary AI systems-particularly machine learning models and large language models-excel at pattern recognition, large-scale document analysis, and probabilistic inference, these capabilities do not translate into reliable adjudicative judgment. Algorithmic systems remain vulnerable to hallucination, opacity, adversarial manipulation, and failures of contextual understanding, and they lack the embodied social reasoning and moral accountability that jury verdicts presuppose. Even if AI systems were eventually able to outperform human jurors in predictive accuracy, replacing juries with algorithmic decision-makers would transform adjudication from a democratic process of collective judgment into a technocratic system of optimized prediction, thereby undermining the institutional legitimacy that jury verdicts provide. The most plausible near-term role for AI is therefore augmentation rather than substitution: AI may assist juries by organizing evidence, reconstructing timelines, and translating complex expert testimony, but the ultimate act of judgment must remain human.