Deceived by ChatGPT Lies... 30-Year Veteran Lawyer Heads to Hearing

Six Submitted Cases Are Fake "Unprecedented"
Lawyer "Didn't Know Because It Was My First Time Using ChatGPT"

An artificial intelligence (AI) chatbot, ChatGPT, presented fictional legal materials in response to a request from an American lawyer to find legal documents. The lawyer faced the risk of being referred to a hearing for submitting documents containing 'false precedents.'

According to a report by The New York Times (NYT) on the 28th (local time), a man named Roberto Mata recently filed a lawsuit against Colombia's Avianca Airlines. He claimed that he was injured in the leg after being hit by a metal cart used for transporting food on an Avianca Airlines passenger plane traveling from El Salvador to New York in August 2019, and argued that it was the fault of the airline staff.

The airline requested dismissal at the Manhattan Federal Court, stating that the lawsuit was filed after the general statute of limitations for airline incidents (2 years) had passed. Steven Schwartz, a lawyer with 30 years of experience representing Mata, insisted, "The lawsuit must proceed." He submitted a 10-page brief to the court containing more than six similar precedents.

Bart Banino, a lawyer for Avianca Airlines who reviewed the brief, sent a letter to Judge Kevin Castel of the Southern District of New York Federal Court, stating, "We could not find the precedent involving China Southern Airlines cited in Schwartz's brief, nor the Korean Air ruling from the 11th Federal Circuit Court of Appeals in 2008 referenced therein." He further raised suspicions that AI chatbots might have been involved, saying, "The precedents included in the brief may not be actual rulings."

[Image source=Yonhap News]

Judge Castel inquired with the 11th Federal Circuit Court of Appeals about the matter and received a response that "a search by the case number provided by Schwartz yielded a completely unrelated case." The other five precedents were also found to be fabricated.

Judge Castel told the NYT, "Schwartz's brief was full of forged fake judicial decisions and fake citations," adding, "This is an unprecedented case in the court." He announced that a hearing would be held on the 8th of next month to discuss whether to sanction lawyer Schwartz.

Lawyer Schwartz admitted that he had used ChatGPT's help to prepare the brief. He stated, "I had no intention to deceive the court or the airline. I had never used an AI chatbot before, so I did not realize that the materials could be fake." He added, "I repeatedly asked ChatGPT to verify whether the precedents existed, and each time it responded, 'It is true. All are real cases and can be found in reputable legal databases,'" appealing for leniency.

The NYT commented, "This case shows that although there are growing concerns that AI threatens the existence of many professional jobs, there is still time before full replacement."

This content was produced with the assistance of AI translation services.