In pre-digital times, a veteran magistrate of the Court of Barcelona left one of those phrases always remembered by court chroniclers.
"Find yourself a lawyer who will defend you from your lawyer," he advised a defendant during the trial.
Today, several decades later, in the nascent world of artificial intelligence (AI), that sentence requires an update, but it remains valid because there are lawyers who continue to be just as illiterate.
You only have to verify this case registered in the New York courts, where it is shown that artificial intelligence has the ability to imitate something as human as traps at a young age.
A citizen named Roberto Mata filed a lawsuit against the airline Avianca. He reported that he was injured when a metal service cart hit him in the knee on a flight to New York's JFK airport. The company asked the federal judge in Manhattan to dismiss the matter and the passenger's lawyer strongly opposed the inadmissibility.
As a supposedly good jurist, attorney Steven S. Schwartz delivered a ten-page report rich in case law.
It incorporated a dozen “important” judicial decisions in which the validity of the complaint and the effect of automatic suspension on the possible prescription were defended.
Among the resolutions he included those of the lawsuits "Martínez against Delta Air Lines", "Zicherman against Korean Airl Lines" and, especially relevant to his argument, "Varghese against China Souther Airlines".
Neither Avianca's lawyers nor the judge himself could find those resolutions or the quotes and references that were incorporated into the document.
The mystery had an explanation. The world famous AI tool called ChatGPT made it all up.
Various tricks are attributed to this mechanism, such as manipulating photographs to win prizes, creating songs as if they were musical stars or writing school papers and speeches. And this is only the beginning.
"At least six of the cases presented appear to be false judicial decisions, with false quotation marks and false internal information," Judge Castel stressed in the order to request clarification from the lawyer. In his letter he alluded to "an unprecedented circumstance."
From lawyer Schwartz, from buffet Levidow, Levidow
Schwartz vowed not to use AI to supplement his inquiries "without absolute verification of its authenticity." It is clear that ChatGPT, the result of his time, goes through the algorithm that "swears to tell the whole truth ...".
