The Expert Record logo — a ruled transcript pageThe Expert RecordThe standard for expert witnesses using AI

AI in the Courtroom · Fabricated citations and facts in filings

Mata v. Avianca, Inc.

Plaintiff's lawyers in a personal injury suit against Avianca filed a brief citing cases produced by ChatGPT that did not exist. Judge P. Kevin Castel imposed a $5,000 penalty and ordered letters to the client and to the judges named as authors of the fake opinions.

Primary document read$5,000 penalty; letters to client and named judges June 22, 2023

What happened

Roberto Mata sued Avianca for an injury he said occurred when a metal serving cart struck his left knee on a 2019 flight. Avianca moved to dismiss. On March 1, 2023, Mata's lawyers at Levidow, Levidow & Oberman filed an affirmation in opposition that cited decisions including Varghese v. China Southern Airlines, Shaboon v. Egyptair and Martinez v. Delta Airlines.

Avianca's counsel could not find the cases. Neither could the court. The research had been done with ChatGPT, and the cited decisions did not exist. When the court ordered copies, Peter LoDuca filed an affidavit on April 25, 2023 that annexed what the court described as "purported" copies or excerpts of the decisions. The court found that LoDuca signed the opposition but that it "was researched and written by Mr. Schwartz," referring to Steven A. Schwartz.

What the court did

In an Opinion and Order on Sanctions dated June 22, 2023, Judge P. Kevin Castel found that the individual lawyers acted in subjective bad faith. The opinion states that they "continued to stand by the fake opinions after judicial orders called their existence into question." It lists the harms of fake citations, including wasted time and money for the opposing party, diverted court resources and damage to the reputation of judges whose names are falsely invoked. The order states: "A penalty of $5,000 is jointly and severally imposed on Respondents and shall be paid into the Registry of this Court within 14 days of this Opinion and Order." The respondents were the two lawyers and the firm.

The court also directed the respondents to send letters to Mr. Mata and to each judge falsely identified as the author of a fabricated opinion, enclosing the sanctions order. The opinion includes the sentence: "Technological advances are commonplace and there is nothing inherently improper about using a reliable artificial intelligence tool for assistance."

In a separate order the same day, the court granted Avianca's motion to dismiss the claim under the two-year limit of the Montreal Convention.

Where it stands

The sanctions order is final for this record. Reporting on later matters, including the Victorian murder case recorded in this section, refers back to it as the 2023 United States case in which a federal judge fined 2 lawyers and a firm over ChatGPT-generated research.

The obligation it engaged

The conduct engaged Clause 3 of the Expert Record Standard (independent authority). Each case entered the brief on the system's authority. None was confirmed in a reporter or database before filing. The later submission of generated "opinions" repeated the same step: the system was asked to verify its own output.

Clause 4 (checkability before adoption) applies directly. A citation to a federal reporter is among the most checkable items in legal work. The governing test is plain here. A competent lawyer with Westlaw, Lexis or a public docket, and no AI system, could not have produced these citations at all.

Sources

What could not be confirmed

Any later appeal of the dismissal or the sanctions order was not checked.

About this entry

This entry was checked against the primary document listed in the sources. The entry for Mata v. Avianca, Inc. records conduct and what the tribunal did. It does not characterize a holding and it is not legal advice. Last checked September 11, 2026. More matters like this one are collected under fabricated citations in filings. If something here is out of date, write to [email protected].

Arrow pointing upTop