AR | Incident Report
Hallucination
2026-01-02
San Luis appeals court flags AI-style hallucinated citations in defense filing
AI Model: Unspecified generative AI tool
I. Executive Summary
Argentine media reported that an Impugnation Tribunal in San Luis rejected a defense appeal after identifying a cited precedent that could not be substantiated and was described as consistent with an AI “hallucination”.
The tribunal reportedly required the lawyer to identify the source details but the lawyer was unable to do so and asked to remove the citation instead.
The reporting indicates the case may be referred for professional review and possible sanctioning.
II. Key Facts
- Impugnation Tribunal (San Luis, First Circumscription) rejected a defense appeal in a sexual-abuse case.
- The filing cited a purported Argentine Supreme Court precedent that the tribunal could not verify.
- The tribunal described the issue as consistent with an AI “hallucination” and requested source details from counsel.
- Counsel reportedly could not provide verifiable details and sought to strike the citation.
- Reporting indicates potential notification to the local bar/professional body for review.
III. Regulatory & Ethical Implications
Highlights growing judicial intolerance for unverified AI-assisted legal authorities and reinforces counsel’s duty to validate citations before filing.
The incident increases sanction and disciplinary exposure (procedural good-faith / professional diligence) and supports adoption of mandatory verification controls (human review, database checks, source retention) for any AI-assisted drafting.