Lawyers suing Colombian airline Avianca give a brief full of past cases that were fair Consisting of by ChatGPT, New York times reported today. After the opposing attorney pointed out the non-existent cases, U.S. District Judge Kevin Castell asserted that “six of the cases presented appeared to be bogus court decisions with bogus citations and bogus internal citations,” and held a hearing as he considered penalties for the plaintiff’s attorney.
Attorney Stephen A. Schwartz in an affidavit that he used OpenAI’s chatbot in his research. To check the cases, he did the only sensible thing: he asked the chatbot if it was a liar.
When asked for a source, ChatGPT apologized for the earlier confusion and insisted the case was real, saying it could be found on Westlaw and LexisNexis. Satisfied, he asked if the other cases were fake, and ChatGPT confirmed that they were all real.
The opposing attorney made the court aware of the matter in excruciating detail as he recounted how brief Livido, Livido, and Opperman’s attorneys’ submission was full of lies. In one example, a non-existent case called Varghese v. China Southern Airlines Co., Ltd. , the chatbot appears to be referring to another state TRUE Zicherman v. Korean Air Lines Co., Ltd. , but got the date (and other details) wrong, saying it was decided 12 years after its original 1996 decision.
Schwartz says he “wasn’t aware that its content could be false”. He now “deeply regrets having used generative AI to supplement the legal research being conducted here and will never in the future do so without absolute validation.”
Schwartz was not admitted to practice in the Southern District of New York, but he originally filed the lawsuit before it was transferred to that court and says he has continued to work in it. Another attorney at the same firm, Peter Loduca, has become the registered attorney for the case, and will have to appear before a judge to explain what happened.
Anyway, here’s the judge pointing out all the ways the lawyer’s brief was an absolute lie: