Mata v. Avianca is the 2023 federal case in which a New York lawyer filed a brief built on six court decisions that did not exist, all generated by ChatGPT, and was sanctioned along with a colleague after they failed to catch the fabrications and then failed to admit them promptly. The underlying dispute was a personal injury claim against an airline. What made the case famous was the collision between a lawyer’s duty to verify sources and a generative AI tool that produced confident, fluent, entirely fictional legal citations.
The Injury Claim Behind the Case
Roberto Mata said a metal serving cart struck his knee during an international Avianca flight in August 2019. He first sued in July 2020, but that case stalled because Avianca was in bankruptcy and protected by an automatic stay. After Avianca emerged from bankruptcy in early 2022, Mata voluntarily dismissed the original complaint and filed a new one on February 2, 2022, in New York state court.1Justia Law. Mata v. Avianca, Inc., No. 1:2022cv01461 – Document 54 (S.D.N.Y. 2023)
Avianca removed the case to the Southern District of New York and moved to dismiss. The airline argued that the Montreal Convention governed the claim, and that treaty imposes a strict two-year filing deadline under Article 35. If a passenger does not bring a claim within two years of arrival, the right to damages is extinguished.2IATA. Convention for the Unification of Certain Rules for International Carriage by Air (Montreal Convention) Mata’s 2022 refiling was well beyond that window.
The Six Fake Cases
Mata’s attorney, Peter LoDuca, filed an opposition brief that cited a series of prior decisions supporting the argument that the case should survive. Six of those decisions were fabricated: Varghese v. China Southern Airlines, Martinez v. Delta Airlines, Shaboon v. Egyptair, Petersen v. Iran Air, Estate of Durden v. KLM Royal Dutch Airlines, and Miller v. United Airlines.1Justia Law. Mata v. Avianca, Inc., No. 1:2022cv01461 – Document 54 (S.D.N.Y. 2023) They had realistic case numbers, plausible court names, and quoted passages from opinions that did not exist.
LoDuca had not drafted the research. His colleague at Levidow, Levidow & Oberman, Steven A. Schwartz, had done that work using ChatGPT. Schwartz had practiced law for more than 30 years and later said he had never used the tool before and did not understand it could produce convincing fictional citations. When he grew uneasy, he went back to ChatGPT and asked whether the cases were real. The chatbot said yes and told him they could be found on Westlaw and LexisNexis. Schwartz never checked either database.3Berkeley Law. Mata v. Avianca, Inc., 678 F.Supp.3d 443 (S.D.N.Y. 2023)
How the Fabrication Came to Light
Avianca’s lawyers could not locate any of the cited decisions. They flagged it to the court, and Judge P. Kevin Castel ordered LoDuca to file copies of the opinions by a set deadline, warning that failure to comply would result in dismissal under Rule 41(b). Instead of acknowledging the problem, LoDuca filed an affidavit attaching purported excerpts from the decisions. Those excerpts were themselves fabricated content from ChatGPT. He also requested a deadline extension, citing a vacation.1Justia Law. Mata v. Avianca, Inc., No. 1:2022cv01461 – Document 54 (S.D.N.Y. 2023)
Judge Castel later noted that the attorneys “doubled down” and did not begin telling the truth until May 25, 2023, after an order to show cause why sanctions should not be imposed. The delay drove much of what followed. An immediate admission when Avianca first raised the alarm would likely have produced a very different outcome for the lawyers personally.
What the Judge Found Wrong With the Opinions
Judge Castel examined the AI-generated decisions and found problems any basic verification would have caught. He described the reasoning in the fabricated Varghese opinion, presented as an Eleventh Circuit ruling, as “gibberish.”3Berkeley Law. Mata v. Avianca, Inc., 678 F.Supp.3d 443 (S.D.N.Y. 2023) Each fake opinion was attributed to real, identifiable federal judges. The invented Miller decision named Judge Barrington D. Parker of the Second Circuit. Petersen was attributed to Judge Reggie B. Walton. Neither judge had any connection to the cases put under their names, and the quotations and holdings were invented.
The Sanctions
Judge Castel found that Schwartz and LoDuca acted in bad faith, citing “acts of conscious avoidance and false and misleading statements to the Court.” He imposed sanctions under Rule 11 of the Federal Rules of Civil Procedure, which requires every filing to rest on a reasonable inquiry into the facts and the law.4Legal Information Institute. Rule 11 – Signing Pleadings, Motions, and Other Papers; Representations to the Court; Sanctions The attorneys’ failure to verify any of the AI-generated citations fell far short of that standard.
The sanctions had several parts:
- A $5,000 penalty imposed jointly and severally on Schwartz, LoDuca, and the firm Levidow, Levidow & Oberman P.C., payable to the court registry within 14 days.1Justia Law. Mata v. Avianca, Inc., No. 1:2022cv01461 – Document 54 (S.D.N.Y. 2023)
- Individual letters to every judge falsely named as the author of one of the six fabricated opinions, each letter including a copy of the sanctions order, the hearing transcript, and the fake opinion attributed to that judge.3Berkeley Law. Mata v. Avianca, Inc., 678 F.Supp.3d 443 (S.D.N.Y. 2023)
- A letter to Mata himself with the sanctions opinion, the hearing transcript, and the affirmation containing the fake citations.
The firm also arranged for outside counsel to run a mandatory continuing legal education program on technological competence and artificial intelligence for its attorneys and staff.3Berkeley Law. Mata v. Avianca, Inc., 678 F.Supp.3d 443 (S.D.N.Y. 2023)
What Happened to Mata’s Injury Case
The sanctions ran alongside the merits question, and Mata lost on the merits. In a separate opinion issued the same day, Judge Castel granted Avianca’s motion to dismiss. The court held that the Montreal Convention’s two-year deadline is not an ordinary statute of limitations but a strict condition precedent to bringing a claim, and it cannot be extended through equitable tolling. Mata’s injury occurred in August 2019 and the operative complaint was filed in February 2022. Even accounting for the bankruptcy stay, the two-year window had closed.2IATA. Convention for the Unification of Certain Rules for International Carriage by Air (Montreal Convention)
The fabricated citations did not cause the dismissal. The claim was likely time-barred no matter what was in the brief. But the collapse of the AI-generated research foreclosed the kind of careful argument on tolling and procedural posture that a competently litigated case might have raised.
Why the Case Became a Landmark
The $5,000 fine was modest. What set Mata v. Avianca apart was timing and facts. In early 2023, generative AI tools were being adopted quickly across industries, and the legal profession was arguing about how to use them. The case gave that argument a concrete, embarrassing example.
The failure was not using AI. It was using AI as a substitute for verification. Schwartz treated ChatGPT like a legal database that retrieves real cases, when it is a text-generation model that produces plausible-sounding output whether or not the underlying material exists. When he asked the tool to confirm the cases were real, it did what it was designed to do and generated a confident, fluent response.
How Courts Responded
The case triggered a wave of judicial reaction. By early 2026, a growing number of federal judges had issued standing orders on AI use in legal filings. Some, including one from Judge Dale E. Ho in the same Southern District of New York, require attorneys to disclose whether they used generative AI in preparing submissions. In late 2023, the Fifth Circuit Court of Appeals proposed a rule requiring attorneys to certify either that no AI was used in drafting filings, or that any AI-generated text had been reviewed for accuracy by a human. The Fifth Circuit decided in June 2024 not to adopt the rule, but the proposal itself showed how seriously courts were treating the issue.
Rule 11 already covered the underlying problem. Lawyers have always been required to conduct a reasonable inquiry before filing anything with a court.4Legal Information Institute. Rule 11 – Signing Pleadings, Motions, and Other Papers; Representations to the Court; Sanctions Available sanctions go well past fines. Under Rule 11(c)(4), courts can strike the offending filing, issue formal reprimands, require participation in educational programs, or refer the matter to state disciplinary authorities. Judge Castel’s $5,000 fine sat at the lighter end of that range, likely because the attorneys’ conduct, while reckless, looked less like a scheme to deceive than a cascading failure to admit a mistake. Mata v. Avianca did not create new law. It showed what happens when existing obligations are ignored.