News

AI-Generated Case Law Citations Under Scrutiny: Lessons from Argentine Judicial Practice

Argentine tribunals have recently addressed a growing concern: legal briefs containing fictitious or erroneous court citations and references, presumably produced through artificial intelligence applications. Decisions from courts in Rosario, General Roca, and Morón establish that while AI assistance in legal drafting is permissible, its deployment without rigorous professional verification violates core ethical duties. These rulings offer valuable guidance on balancing technological innovation with the integrity of legal practice.

Argentine judicial authorities have recently confronted a phenomenon previously observed mainly in courts abroad: the submission of legal briefs citing court decisions that, upon verification, proved to be fabricated or materially inaccurate—errors strongly indicative of AI-generated content. These developments signal that what was once primarily a concern in North American courtrooms has now reached Latin America, prompting Argentine judges to articulate standards governing lawyers' professional accountability when leveraging AI technologies.

Three distinct cases across multiple jurisdictions brought this issue: proceedings before the Civil, Commercial, Family, and Mining Appeals Chamber of General Roca; Chamber II of the Civil and Commercial Court of Rosario; and Chamber I of the Civil and Commercial Court of Morón. Each tribunal identified submissions referencing non-existent rulings or containing quotations that did not correspond to any identifiable judicial decision. In several instances, exhaustive searches across court cases databases yielded no matches whatsoever.

The procedural responses varied. In the Giacomino matter, the Rosario appellate court initiated verification measures regarding the cited authorities and demanded an explanation from counsel, who openly acknowledged employing AI to construct and bolster his legal arguments. In M.J.L., comparable procedural steps were taken, yet the attorneys involved declined to confirm or deny AI usage in preparing the contested filings. Meanwhile, in Acevedo, although the court did not formally demand clarification, it was inferred from the distinctive drafting patterns that AI tools had been utilized.

Across all three decisions, the judges concurred on a fundamental principle: utilizing AI as a drafting aid is not inherently improper, but doing so without appropriate professional supervision is unacceptable. This consensus ignited discussion regarding the boundaries of professional liability in an era of AI-assisted legal work. The courts emphasized that filing briefs containing non-existent precedents—even absent malicious intent—contravenes foundational professional obligations including truthfulness, honesty, loyalty, and procedural good faith, all enshrined in the ethical codes governing Argentine legal practice.

Notably, the courts refrained from imposing direct disciplinary measures on the implicated attorneys. However, they determined that notification to the respective bar associations was warranted—both to alert the profession to the inherent risks and potential liabilities associated with AI usage and to stimulate institutional dialogue on responsible implementation of these technologies in legal practice.

A recurring theme throughout these rulings was the distinction between AI utilization itself and its application without subsequent verification. The judges acknowledged AI's undeniable utility as a support mechanism while stressing that these systems are known to "hallucinate"—generating plausible-sounding but entirely fictitious information—cite sources inappropriately or produce imprecise references. Given that such limitations are now extensively documented in both academic literature and professional forums, the standard of diligence expected from practitioners necessarily increases.

The courts articulated an unambiguous position: technological tools cannot substitute for the intellectual labor inherent in legal advocacy, nor can they relieve attorneys of their obligation to verify the accuracy of every element within their submissions. The foreseeable nature of AI-generated errors requires lawyers, now more than ever, to implement robust verification protocols. Recommended practices include confirming each citation independently, maintaining documentary records, and providing hyperlinks to facilitate judicial review.

These Argentine precedents carry significance beyond national borders. Together with similar rulings emerging from jurisdictions worldwide, they demonstrate that the practical implications of AI in legal practice are no longer speculative or limited to technologically advanced legal systems. Simultaneously, they reinforce that while AI represents a valuable professional resource, its responsible use demands consistent human oversight to guarantee the accuracy and integrity of legal submissions.

The cases examined include: "M.J.L. v. Peugeot Citroën Argentina S.A. et al. (summary proceeding)," "Giacomino, César Adrián et al. v. Monserrat, Facundo Damián et al. (damages)," and "Acevedo, Gerardo Gabriel v. Cáceres Mareco, Willian Arsenio (motor vehicle damages with injury or death)."

 

Article provided by INPLP member: Diego Fernandez (Marval, O'Farrell & Mairal, Argentina)

 

 

Discover more about the INPLP and the INPLP-Members

Dr. Tobias Höllwarth (Managing Director INPLP)