World Journal of Law, Governance and Public Policy
Hallucinated Justice; AI-Generated Case Law and the Attorneyâs Obligation to Verify
Abstract
Andrew Ndubisi Ucheomumu
The proliferation of generative artificial intelligence (“AI”) in legal practice has produced a novel and troubling category of professional responsibility violations. Courts across the United States have confronted an increasing volume of pleadings and briefs containing legal authorities that either do not exist or that AI systems have materially mischaracterized. Unlike ordinary citation errors attributable to research lapses, AI-generated hallucinations can produce fabricated case names, fictitious reporters, invented quotations, and nonexistent holdings, all rendered with the superficial confidence of authentic legal authority.
Generative AI is an important tool in today’s legal practice, and it is neither the first nor the last technological innovation to transform the practice of law. From the printing press to the typewriter, from Westlaw's computerized databases to electronic filing systems, each era's defining technology has enhanced attorney productivity, expanded access to legal services, and reshaped the mechanics of legal work. Generative AI is the latest and most powerful entry in that lineage — an indispensable tool that enables attorneys to research, draft, and analyze with a speed and breadth previously unattainable. Critically, the courts that have imposed sanctions for AI misuse have been uniform on one point: they do not condemn the use of AI in legal research and briefing. From Mata v. Avianca to Mezu v. Mezu, the judicial message has been consistent — AI is a legitimate and valuable instrument of legal practice. What the courts condemn, without exception, is the abdication of professional judgment that occurs when attorneys submit AI-generated work product without independent verification. The obligation is not to avoid AI, but to use it as every transformative tool before it has always demanded to be used: with skill, with diligence, and with the lawyer's own professional accountability firmly intact.
This Article surveys the rapidly developing body of federal and state case law addressing sanctions for the submission of AI-generated fabricated citations and for AI misrepresentation of authentic precedent. Beginning with the foundational decision in Mata v. Avianca, Inc., the Article traces the emergence of a national sanctions framework across numerous federal circuits and district courts, as well as courts in California, Hawaii, Louisiana, Maryland, and Utah. It examines the distinct but related problem of AI systems that accurately cite real cases but misrepresent their holdings, a subtler error that may escape initial detection yet pose equal risks to the integrity of adjudication.
The Article then analyzes the ethical obligations that existing professional responsibility rules impose on attorneys who use AI tools, including the duties of competence, diligence, and candor toward the tribunal, as amplified by American Bar Association Formal Opinion 512. Against this backdrop, the Article examines Maryland’s existing rules and the significance of Mezu v. Mezu (2025) as Maryland’s first published appellate opinion addressing AI-citation misconduct. The Article further examines the emerging question of law firm governance and institutional responsibility, analyzing the extent to which law firms bear independent obligations to implement AI oversight protocols under existing supervisory ethics rules. The Article concludes with concrete recommendations for Maryland courts, bar authorities, and the legislature: that Maryland should adopt AI-certification requirements analogous to those already implemented in several federal districts, should enforce its existing procedural and ethical rules vigorously in cases involving AI-generated errors, should issue definitive bar guidance to assist practitioners in navigating the responsible use of generative AI tools in legal practice, and should enact statutory authority enabling appellate courts to impose monetary sanctions sua sponte, closing the enforcement gap that constrained the court’s response in Mezu v. Mezu.