•  
  •  
 
Journal of Technology Law & Policy

Document Type

Article

Abstract

Generative artificial intelligence has entered legal practice with unprecedented speed, transforming how lawyers and judges synthesize information, develop arguments, and evaluate evidence. While early attention has focused on visible failures such as hallucinated citations and fabricated authority, these incidents do not capture the deeper structural risk AI poses to the legal system. The authors argue that generative AI is a stress test for legal institutions: it offers substantial opportunities to improve legal practice and access to justice, but it also operates as a force multiplier that can deepen structural asymmetries when professional competence and oversight are uneven. Unlike prior legal technologies that primarily assisted retrieval or organization, generative systems participate directly in synthesis, framing, and analysis, altering how legal knowledge is formed and contested. Where AI competence is uneven, disparities in analytical capacity, litigation leverage, and epistemic control emerge in ways existing doctrine was not designed to absorb.

Drawing on the historical evolution of electronic discovery, Technology-Assisted Review, and digital evidence, this Article demonstrates why prior models of incremental adaptation are no longer sufficient. Ethical doctrine and sanctions regimes can identify individual failures, but they operate reactively and cannot correct structural asymmetries. Empirical evidence suggests, however, that the competence divide is not fixed: with appropriate education, supervision, and institutional design, generative AI can narrow disparities rather than entrench them. This Article concludes by proposing a calibrated institutional roadmap focused on formation, oversight, incentives, and coordination to ensure that AI advances the core commitments of fairness, accuracy, and legitimacy in an adversarial system increasingly shaped by probabilistic tools.

Download

Share

COinS