Published on: 11 June 2023 at 19:06 IST
In a recent incident, a US lawyer used the ChatGPT artificial intelligence program to prepare a court filing, but the program generated fake cases and rulings, resulting in an embarrassing situation for the attorney.
Steven Schwartz, a lawyer based in New York, apologized to a judge after submitting a brief filled with falsehoods that were created by the OpenAI chatbot.
Schwartz expressed his lack of awareness regarding ChatGPT’s ability to fabricate complete case citations and judicial opinions, particularly in a manner that seemed authentic. This apology was stated in a court filing.
In a civil case at Manhattan federal court involving a man’s lawsuit against Avianca, a Colombian airline, an embarrassing mistake occurred. Lawyer Steven Schwartz submitted a response to the court, generated by the OpenAI chatbot, containing fabricated case citations and judicial opinions.
The case involved Roberto Mata, who claimed to have been injured by a metal serving plate during an Avianca flight in 2019. In Schwartz’s response to the airline’s motion to dismiss the case, he cited several non-existent cases, such as Petersen v. Iran Air, Varghese v. China Southern Airlines, and Shaboon v. Egyptair.
Even internal citations and quotes were included in the Varghese case. The airline’s lawyers and the presiding judge, P. Kevin Castel, could not find any trace of these cases. Schwartz was compelled to admit that the chatbot had generated everything. Judge Castel deemed it an unprecedented situation, with six of the cited cases being fabricated with bogus quotes and citations.
The judge ordered Schwartz and his law partner to appear before him to potentially face sanctions. In a recent court filing, Schwartz expressed deep regret and apologized to the court for the mistake, stating that he believed ChatGPT was a reliable search engine at the time. He insisted that he had no intention to mislead the court.
The incident has garnered media coverage, leading to public ridicule for Schwartz and his law firm, Levidow, Levidow & Oberman. Schwartz acknowledged the deep embarrassment caused by the situation, both personally and professionally, as the articles would be available for an extended period. The story was first reported by The New York Times, and there has been no immediate comment from OpenAI on the matter.