A lawyer used ChatGPT to do "legal investigation" and cited a number of nonexistent cases in a filing, and is now in a lot of trouble with the judge
pic.twitter.com/AJSE7Ts7W7— Daniel Feldman (@d_feldman) May 27, 2023
As The New York Times reports , after opposing counsel presented the non-existent cases, U.S. District Judge Kevin Castel responded forcefully.
He said that six cases have been identified that appear to be false court decisions, with falsified internal references and citations. In addition, the judge called a hearing for June 8 to analyze possible sanctions for the plaintiff's lawyers.
Attorney Steven A. Schwartz uses ChatGPT as an investigative tool
In his affidavit, attorney Steven A. Schwartz acknowledged using OpenAI's chatbot as an investigative tool . In order to verify the veracity of the cases, he proceeded to ask the chatbot directly if it was lying, without checking its veracity.
When asked for a source, ChatGPT regretted the earlier confusion and emphatically stated that the case in question was real , assuring that it could be found on Westlaw and LexisNexis .
Satisfied with this answer, the lawyer proceeded to ask about the other cases, to which ChatGPT forcefully stated that all of them were authentic .
Opposing counsel presented the court with a detailed account of the presentation by Levidow, Levidow & Oberman, noting that it was riddled with falsehoods.
In the lawyer's defense, he submitted screenshots FROM CHATGPT claiming that the nonexistent cases exist pic.twitter.com/H2tgXu8W5s
— Daniel Feldman (@d_feldman) May 27, 2023
Among the examples presented was a non-existent case called Varghese v. China Southern Airlines Co., Ltd. Interestingly, ChatGPT referenced another real case, Zicherman v. Korean Air Lines Co., Ltd., but provided incorrect information, such as the date of the decision, wrongly stating that it was made 12 years after its original ruling in 1996, as TheVerge claims .
Schwartz, the accused lawyer, claims that he was “unaware of the possibility that its jamaica number screening content could be false .” He admits that it was the first time he had used this type of tool and he was unaware that it was not authentic. He swore that he regrets having used this artificial intelligence to complement the investigation and that he will not do so again without verifying whether the information is authentic.
The lawyer stressed that he had no intention of misleading the court and completely exonerated another lawyer from the firm who also exposes himself to possible sanctions.
lawyer uses chatgpt
Are ChatGPT and other artificial intelligence reliable?
Microsoft’s debut on Bing has now acquired an infamous association with outright lies, deception, and emotional manipulation. In its first demo, Google’s AI chatbot Bard even made up a fact about the James Webb Space Telescope .
For example, Meta's chatbot repeats fake news and anti-Facebook comments.