Facts
Roberto Mata sued the airline Avianca in New York state court (Supreme Court of the State of New York, New York County). On February 22, 2022 Avianca removed the case to the U.S. District Court for the Southern District of New York. Avianca moved to dismiss, arguing that the claims were time-barred under the Montreal Convention. The opposition filed by Mata’s lawyers on March 1, 2023 cited several judicial decisions that did not exist; one of the lawyers had generated them with ChatGPT. When Avianca’s counsel and the court could not locate the cases, the lawyers did not withdraw the brief and later filed an affirmation attaching purported copies of the fake opinions. After orders to show cause, the court held a sanctions hearing on June 8, 2023.
Question
The court had to decide whether the lawyers and their firm had violated Rule 11 of the Federal Rules of Civil Procedure, which makes the attorney who signs a filing responsible for ensuring that its legal contentions are warranted by existing law, and whether sanctions should be imposed under that rule or the court’s inherent power. A related question was whether the individual lawyers had acted in subjective bad faith. That turned on their conduct after the existence of the cited cases was first called into question.
Decision
Judge P. Kevin Castel imposed sanctions. He noted that there is nothing inherently improper about using a reliable AI tool for assistance, but that existing rules give attorneys a gatekeeping role over the accuracy of their filings. The lawyers abandoned that role when they submitted fake opinions and kept standing by them after the court questioned their existence. The court found bad faith on the part of the individual lawyers and held the firm jointly and severally liable. It ordered them to send letters, with the opinion attached, to their client and to each judge falsely named as the author of a fake opinion, and to pay a $5,000 penalty into the court’s registry. It declined to compel an apology.
Why it matters
Mata v. Avianca is the best-known early decision on fabricated AI citations. The court found no fault in the use of AI as such. It faulted the failure to verify the citations and the lack of candor once the problem surfaced. Because the reasoning rests on existing duties of diligence and candor, it applies regardless of which tool produced the error.