Amparo Trejo v. Amaya Hernandez (Maryland 2026): Appellate Court Opinion Itself Contains AI-Hallucinated Citations | Advocate Prakhar

Case at a Glance

Full Citation	Amparo Trejo v. Miguel Angel Amaya Hernandez et al., 2026 WL 1162579, No. 0741, September Term, 2025 (Appellate Court of Maryland, April 29, 2026)
Court	Appellate Court of Maryland
Date	April 29, 2026
AI Tool / Issue	The court’s own published opinion contains citation references that are incorrect or do not exist — a new dimension where AI hallucinations enter the judicial opinion rather than just parties’ briefs
Outcome	Appeal dismissed; invalid citations preserved in official record; Westlaw Editor’s Note warns researchers about the citation problems in the opinion

Background

Appellant Amparo Trejo, the aunt of an El Salvador native who entered the US in 2021 to escape gang violence, filed a petition for custody and Special Immigrant Juvenile Status (SIJS) predicate findings for her niece. The circuit court denied an emergency motion to expedite before the minor turned 21 (SIJS eligibility ends at 21). Trejo then filed a Motion for Order Nunc Pro Tunc seeking equitable relief, which was also denied. Trejo appealed.

The Appellate Court of Maryland addressed the appeal and ultimately dismissed it. The Westlaw publication carries the Editor’s Note warning: “This decision contains discussion of citation references that are incorrect or do not actually exist. These invalid citations appeared in the original court opinion and have been preserved as written since they are part of the official record.” This means the AI hallucination problem is within the court’s opinion itself — not in the parties’ briefs — a qualitatively different and more concerning development.

The AI Issue

The Amparo Trejo Editor’s Note represents the most alarming development in the AI hallucination landscape to date: AI-generated hallucinated citations appearing in the text of an official appellate court opinion. When AI hallucinations appear in parties’ briefs, the court can catch them and impose sanctions. When they appear in the court’s own published opinion, they become part of the authoritative legal record — potentially cited by future courts and practitioners as valid precedents supporting propositions they do not actually stand for.

What the Court Decided

• The appeal was dismissed on substantive grounds; the SIJS nunc pro tunc argument was rejected.
• Invalid citations in the court’s opinion were preserved as part of the official record and flagged by Westlaw’s Editor’s Note to warn future researchers.
• The case illustrates that AI use in judicial opinion drafting — by law clerks, court staff, or through AI-assisted judicial writing tools — creates the same hallucination risks as AI use by parties and attorneys.

“This decision contains discussion of citation references that are incorrect or do not actually exist. These invalid citations appeared in the original court opinion and have been preserved as written since they are part of the official record. Any links to these invalid citations have been removed.”

— Westlaw Editor’s Note, Amparo Trejo v. Amaya Hernandez, April 29, 2026

The India Angle

Indian Law Equivalent

Indian judicial opinions are authored by judges with assistance from law clerks and court staff. The Supreme Court and High Courts have not yet issued formal guidelines on AI use in judicial opinion drafting. If AI hallucinations were to appear in a published Indian Supreme Court judgment, the usual remedy would be a review petition under Article 137 of the Constitution or a clarification order — but the hallucinated citation would be part of the official record during the period before correction. The Law Commission of India has noted the need for AI governance in the judicial system but has not yet addressed AI use in opinion drafting specifically.

Bar Council Rules

This case raises issues beyond advocate conduct — it concerns judicial use of AI. While the BCI Rules govern advocates, not judges, the broader implication is that the entire legal system needs AI governance frameworks. Advocates who discover that a relied-upon precedent contains AI-hallucinated citations should bring this to the court’s attention through a formal note, creating a record of the accuracy concern.

Practical Advice for Indian Advocates

• When a precedent you intend to cite is published with an Editor’s Note warning about citation errors, verify independently that the specific proposition you are relying on is actually supported in the opinion text.
• If you discover that a court’s published opinion contains what appear to be AI-hallucinated citations, consider filing a brief noting the concern — courts appreciate accurate records and may issue a corrigendum.
• As AI use in judicial drafting grows, verify judicial quotations with the same rigor as party submissions — the source of a hallucination does not change its impact on the accuracy of the legal record.

Quick Takeaways

• AI hallucinations can enter judicial opinions themselves — not just party briefs — creating unreliable precedents.
• Westlaw’s Editor’s Note system flags these issues to prevent downstream reliance on hallucinated court citations.
• Any citation to a flagged opinion requires independent verification of the relevant propositions.

---

Deep Dive: When AI Hallucinations Enter Judicial Opinions

The Amparo Trejo Editor’s Note reveals a dimension of AI risk that transcends the sanctions framework entirely. In all other AI hallucination cases, the problem is that parties submitted inaccurate citations to courts — and courts can sanction, warn, and disregard those inaccurate submissions. But when the court’s own published opinion contains AI-hallucinated citations, the institutional response is qualitatively different. The court cannot sanction itself; the opinion is part of the official record; and Westlaw’s only option is to add a warning note and remove the hyperlinks to the nonexistent citations.

How does an AI hallucination enter a court opinion? Courts increasingly use AI-assisted research and drafting tools in their law clerks’ workflows. A law clerk using a generative AI tool to draft an opinion passage — particularly in a rapidly-evolving area of law like immigration status proceedings — might include AI-generated citations without independently verifying them. If the judge reviews the opinion for reasoning quality and clarity but does not verify every citation against primary sources, hallucinated citations can survive into the published opinion.

The precedent contamination risk is serious. If Amparo Trejo is cited by a future Maryland court or attorney for the propositions supported by its AI-hallucinated citations, those hallucinations propagate through the legal record — each successive citation making the underlying fabrication harder to trace and correct. Westlaw’s Editor’s Note system helps by flagging the opinion for researchers, but this system depends on researchers actually reading the warning and independently verifying before relying on the opinion.

For the Indian judiciary, this case is a warning bell. The Supreme Court of India and the High Courts are beginning to use AI tools for research and administrative purposes. If AI hallucinations enter published Indian court decisions — even unreported ones cited in subsequent proceedings — the contamination of Indian legal precedent becomes a systemic risk. The appropriate response is for courts to adopt verification protocols for AI-assisted opinion drafting that require independent confirmation of every citation before the opinion is published, just as we ask advocates to do before filing their briefs.