Claiming that he was injured when a metal serving cart struck his knee during a flight from El Salvador to New York in 2019, Robert Mata recently sued Avianca Airlines. Avianca filed a motion to dismiss in New York federal court arguing the lawsuit was too late; the statute of limitations had expired. Vehemently objecting, Mata’s lawyers filed a 10-page reply brief citing more than half a dozen apparently-relevant court decisions. Among them was Varghese v. China Southern Airlines which purported to offer a learned discussion of federal law and “the tolling effect of the automatic stay on a statute of limitations.” What if none of Mata’s reply brief was true?

Avianca and the Judge’s Response

Neither Avianca’s lawyers nor the trial court judge could find the decisions or the quotations cited and summarized in the Mata brief. ChatGPT had invented everything. The plaintiff’s lawyer Steven Schwartz cut and pasted the ChatGPT “brief.” Schwartz, a lawyer for over thirty years, finally confessed that he had used the artificial intelligence program to do his legal research and admitted that it was “a source that has revealed itself to be unreliable.” Insisting that he had no intent to deceive the court or the airline, he offered that he had even asked the program to verify that the cases were real – “[I]t had said ‘yes.’” His excuse? He had never used ChatGPT, and “therefore was unaware of the possibility that its content could be false.” Schwartz said he “greatly regrets” relying on ChatGPT “and will never do so in the future without absolute verification of its authenticity.” Can a lawyer just cut and paste a ChatGPT report into their court filings with impunity? Shouldn’t Schwartz have had clear-cut verification of the citations and verifications before he used the ChatGPT report?

Federal Law and Texas State Law

Yes. Both federal law and Texas state court law require every pleading, written motion and “other paper” filed with a court to be signed by the party or their attorney. In federal court, the signed pleading certifies that, to the best of the person’s knowledge, information and belief, formed after an inquiry reasonable under the circumstances, the claims, defenses and other legal contentions are warranted by existing law. A Texas signed pleading is a certificate by the lawyer that, to the best of their knowledge, information and belief formed after reasonable inquiry, the instrument is not groundless. Failure to abide by these rules can result in sanctions. Allowing that he had been presented with “an unprecedented circumstance” – a reply brief replete with “bogus judicial decisions, with bogus quotes and bogus internal citations” – the judge ordered a hearing to discuss potential sanctions.

The Likely Result?

Who knows? Schwartz, the brief “writer,” threw himself on the mercy of the court with an affidavit that the artificial intelligence program used to do his legal research was “unreliable.” But, was a lack of intent to deceive the court or the airline sufficient to avoid sanctions? I think not. Schwartz, as with opposing counsel and the judge, had a duty before filing his reply brief to form an independent opinion that his position was supported by existing law – real law. How about Schwartz’s “great regret” and promise to “never do so in the future without absolute verification of its authenticity?” Likewise, Schwartz’s signature certified that he had made “reasonable inquiry” and that his reply brief was not groundless. Well, it was… groundless. Moreover, it was bogus. Perhaps Schwartz will get a pass on this one because of his tenure and his general unfamiliarity with how ChatGPT works. Who knows?

Tilting the Scales in Your Favor

If you are a lawyer, beware… and be aware. Given the clearly stated duty of every lawyer to double check and be willing to “stand behind” what is presented in a brief or other court motion, I seriously doubt whether many judges will be forgiving of such “mistakes” – even less likely, federal court judges.

Real Life

What does this apparent lawyer snafu have to say to the rest of us? Should we each have a standing practice to “trust but verify” whenever information – gossip or internet – is repeated and it just doesn’t sound right? Too good to be true? Deviates from “common sense?” And, the like. Is it appropriate to ask for the back-up data? Should we take the time? Is it worth it not to take the time? What might the implications be if all of us fail to take the time?

As a general rule, whether you are a lawyer or not, you should always verify the information ChatGTP creates. It can be an excellent tool for drafting and finding information, but don’t depend on it to be accurate. Always double check your work.

Attending a wedding in upstate New York, my son, Cam Clinton, read and shared with me this story first reported in the NY Times over the 2023 Memorial Day weekend. Stay tuned for the “rest of the story.”

A Southern Solution

Gray Reed Associate Donte Jones apprises me that, apparently mindful of the ChatGPT pitfalls, Northern District of Texas Judge Brantley Starr now requires all attorneys appearing before his Court to file a docket certification attesting either that no portion of any filing will be drafted by generative artificial intelligence, or that any language drafted by artificial intelligence will be checked for accuracy by a human being. For those Texas attorneys, does this new requirement impact FRCP 11? Stay tuned.

[View source.]