The lawsuit began like so many others: A man named Roberto Mata sued the airline Avianca, saying he was injured when a metal serving cart struck his knee during a flight to Kennedy International Airport in New York.
When Avianca asked a Manhattan federal judge to toss out the case, Mata’s lawyers vehemently objected, submitting a 10-page brief that cited more than half a dozen relevant court decisions. There was Martinez v. Delta Air Lines, Zicherman v. Korean Air Lines and, of course, Varghese v. China Southern Airlines, with its learned discussion of federal law and “the tolling effect of the automatic stay on a statute of limitations”.
There was just one hitch: No one — not the airline’s lawyers, not even the judge himself — could find the decisions or the quotations cited and summarised in the brief. That was because ChatGPT had invented everything.
The lawyer who created the brief, Steven A. Schwartz of the firm Levidow, Levidow & Oberman, threw himself on the mercy of the court on Thursday, saying in an affidavit that he had used the artificial intelligence program to do his legal research — “a source that has revealed itself to be unreliable”. Schwartz told Judge P. Kevin Castel that he had no intent to deceive the court or the airline. Schwartz said that he had never used ChatGPT, and “therefore was unaware of the possibility that its content could be false”. He had, he told Judge Castel, even asked the program to verify that the cases were real. It had said yes.
Schwartz said he “greatly regrets” relying on ChatGPT “and will never do so in the future without absolute verification of its authenticity”. Judge Castel said in an order that he had been presented with “an unprecedented circumstance”, a legal submission replete with “bogus judicial decisions, with bogus quotes and bogus internal citations”. He ordered a hearing for June 8 to discuss potential sanctions. As artificial intelligence sweeps the online world, it has conjured dystopian visions of computers replacing not only human interactions but also human labour. The fear has been especially intense for knowledge workers, many of whom worry that their daily activities may not be as rarefied as the world thinks — but for which the world pays billable hours. The real-life case of Roberto Mata v. Avianca Inc. shows that white-collar professions may have at least a little time left before the robots take over. In a brief filed in March, Mata’s lawyers said the lawsuit should continue, bolstering their argument with references and quotes from the many court decisions that have since been debunked.
Soon, Avianca’s lawyers wrote to Judge Castel, saying they were unable to find the cases that were cited in the brief. When it came to Varghese v. China Southern Airlines, they said they had “not been able to locate this case by caption or citation, nor any case bearing any resemblance to it”. They pointed to a lengthy quote from the purported Varghese decision contained in the brief. “The undersigned has not been able to locate this quotation, nor anything like it in any case,” Avianca’s lawyers wrote. Indeed, the lawyers added, the quotation, which came from Varghese itself, cited something called Zicherman v. Korean Air Lines Co. Ltd., an opinion purportedly handed down by the US Court of Appeals for the 11th Circuit in 2008. They said they could not find that, either.
Judge Castel ordered Mata’s attorneys to provide copies of the opinions referred to in their brief. The lawyers submitted a compendium of eight; in most cases, they listed the court and judges who issued them, the docket numbers and dates. The copy of the supposed Varghese decision, for example, is six pages long and says it was written by a member of a three-judge panel of the 11th Circuit. But Avianca’s lawyers told the judge that they could not find that opinion, or the others, on court dockets or legal databases.
In Mata’s case, ChatGPT appears to have discerned the labyrinthine framework of a written legal argument but has populated it with names and facts from a bouillabaisse of existing cases. On Thursday, Mata’s lawyers offered affidavits containing their version of what had happened. Schwartz said that he had consulted ChatGPT “to supplement” his own work and that, “in consultation” with it, found and cited the half-dozen nonexistent cases. He said ChatGPT had provided reassurances. “Is varghese a real case,” he typed, according to a copy of the exchange that he submitted to the judge. “Yes,” the chatbot replied, offering a citation and adding that it “is a real case”.
Schwartz dug deeper. “What is your source,” he wrote, according to the filing. “I apologise for the confusion earlier,” ChatGPT responded, offering a legal citation. “Are the other cases you provided fake,” Schwartz asked. ChatGPT responded, “No, the other cases I provided are real and can be found in reputable legal databases.” But, alas, they could not be.
New York Times News Service
