Page 30 - ACTL_Fall23

In his written submission, Namesake explained himself initially by say- ing he did check the sources referenced in the legal argument generated by ChatGPT. How? He asked ChatGPT. He actually submitted screen shots of his question and answer session with the AI Program:
29
JOURNAL
The eight case reports Mr. Mope filed with the Court look odd, unlike what one would see from Westlaw or LexisNexis. They looked odd be- cause none of these case reports was real. Instead, they were created by the AI program from bits and pieces of other cases and cobbled together into something resembling (very loosely) a real case report; at least that’s what Mr. Mata’s lawyers later claimed/confessed.
Varghese, the Eleventh Circuit case principally relied upon to address the statute of limitations issue, lists the names of three judges who supposedly heard this appeal. I was able to identify two as actual Eleventh Circuit Judges and the third as a Fifth Circuit Judge. The Varghese case report sub- mitted to the judge is full of anomalies and inconsistencies. The decision is undated. The initial paragraph identifies Susan Varghese, the personal representative of the Estate of George Varghese, who has filed a wrongful death claim. The next paragraph refers to the plaintiff as “Anish Varghese” and later paragraphs describe the cause of action as breach of contract. The case report first recites that the defendant airlines filed for bankruptcy in China, but then recites that it was the plaintiff who filed for bankruptcy under U.S. law, though the report first recites that it was under Chapter 7, but later refers to a Chapter 13 filing.
Peterson, a second case report submitted to the court, purports to be an opinion by Judge Walton of the District of Columbia. Judge Walton in- deed sits in that district, but the alleged case report has obvious flaws, such as that the judge grants the motion to dismiss in the first paragraph and denies the same motion in the final paragraph. A third case report, Durden, allegedly from the Georgia Court of Appeals from 2017, does involve a suit against an airline, but recites that the plaintiff was killed in an accident at Hartsfield-Jackson Atlanta International Airport “when he was struck by a tow bar that had detached from a baggage tug ownedand [sic] loss of consortium claims.” A colossal failure of proofreading, it defies credulity that any real court would issue this opinion.
You get the point. So did the judge who has Mr. Mata’s case. After Mr. Mope submitted these obviously bogus case reports, the judge issued a second Show Cause Order requiring Mr. Mope to appear for a hearing as to why sanctions should not be imposed against him. Mr. Mope sought another extension and explained that it was actually Mr. Namesake who had drafted the legal argument submitted to the Court as well as gathering the case reports that were submitted.
Thus a third Show Cause Order, requiring Namesake to also appear at the hearing. In a sign of things to come, the judge refused to grant Mope any further extensions, but did give Namesake two extra days to submit something in writing in advance of the hearing.
Namesake: “Is Varghese a real case?”
ChatGPT: “Yes, Varghese v. China South Airlines Co Ltd, 925 F.3d 1339 (11th Cir. 2019) is a real case.”
Namesake and his firm hired counsel. Namesake’s counsel wrote: “ChatGPT even assured him the cases were real and could be found on Westlaw and LexisNexis, and continued to provide extended excerpts and favorable quotations. Now that Mr. Name- sake and the Firm know ChatGPT was simply making up cases, they are truly mor- tified.” The memo also referenced several articles warning of the dangers of errors that can be made by ChatGPT and the inade- quacy of the warnings that tell users: “May occasionally generate incorrect information.”
Do you think?
At the sanctions hearing, Judge Castel forced Namesake to admit he didn’t do anything to check the AI program’s re- search output. “ChatGPT was not sup- plementing your research. It was your research, correct?” the judge asked.
On June 22, 2023 Judge Castel issued two decisions. The first unsurprisingly granted Avianca’s motion to dismiss. The beginning of the second considerably longer decision is worth quoting:
In researching and drafting court sub- missions, good lawyers appropriately obtain assistance from junior lawyers, law students, contract lawyers, legal encyclopedias and databases such as Westlaw and LexisNexis. Technological advances are commonplace and there is nothing inherently improper about us- ing a reliable artificial intelligence tool for assistance. But existing rules impose a gatekeeping role on attorneys to en- sure the accuracy of their filings. [The lawyers in this case] abandoned their responsibilities when they submitted non-existent judicial opinions with fake quotes and citations created by the ar-