Skip to content
Submit incident
AII-1002-62969
ConfirmedA court, a regulator or the organization itself has stated the core facts, or the record cites a court or regulator source.Supporting sources are pending verification.

Lawyer Sanctioned for ChatGPT-Generated Fabricated Case Citations in Frontier Airlines Appeal

A Maryland attorney was sanctioned $1,000 by the U.S. Court of Appeals for the Tenth Circuit and referred to disciplinary authorities after using ChatGPT to draft an appellate brief that contained seven fabricated case citations and misattributed quotations. The court found the attorney's failure to verify citations constituted reckless disregard of her duty to the court, though she stated she had no intent to misrepresent the law.

24 January 2025 Β· Kusmin L. Amarsingh Β· Reliability and hallucination
24 January 2025
AII-1002-62969
Quick facts
Category
Reliability and hallucination
Country
United States
Date
24 January 2025
Drafted by
Raidu Incident Detection

What happened

Kusmin L. Amarsingh used ChatGPT to research and draft an appellate brief that she filed with the U.S. Court of Appeals for the Tenth Circuit. The brief contained seven fabricated case citations and misattributed propositions and quotations to actual cases.

Upon review, the Tenth Circuit identified the fabricated citations and ordered Amarsingh to explain and correct the filing. She admitted she had used ChatGPT to generate the false citations and stated she had no intention of misrepresenting the law or the record.

Reported impact

The Tenth Circuit found that Amarsingh's actions constituted reckless disregard of her fundamental duty to the court to verify that all cited cases exist and accurately represent the propositions attributed to them. The court determined that careless use of AI tools can waste judicial resources and the opposing party's time and money, and can damage the credibility of the legal system.

Index analysis

The Tenth Circuit sanctioned Amarsingh under Federal Rule of Appellate Procedure 38, which does not require proof of subjective bad faith but examines whether conduct objectively manifests intentional or reckless disregard of duties to the court. The court concluded that while there is nothing inherently problematic with using generative AI in legal practice, Amarsingh's complete failure to perform basic verification duties was reckless.

Governance controls

The Tenth Circuit referred a copy of its order and judgment to attorney-disciplinary authorities in Maryland, where Amarsingh is licensed. Amarsingh completed a continuing legal education course on the ethics of using artificial intelligence after the court proceedings.

Classification

Category
Reliability and hallucination
Country
United States
Harm status
Actual harm
Intent
Unintentional
Timing
Post-deployment
Harm types
Economic or property, Reputational
Failure modes
Fabricated citation, Hallucination, Overreliance
Consequences
Fine or settlement
AI systems
ChatGPT

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident