Artificial intelligence? Not in the courtroom.
Hadera Magistrates Court Judge Ehud Kaplan was shocked to discover that a legal brief submitted by the police in a case included laws that had been fabricated by the artificial intelligence tool ChatGPT.
During a hearing regarding the confiscation of a phone belonging to a suspect under investigation by the Lahav 433 unit, the suspect’s attorney demanded that the device be returned to his client.
The police submitted a legal brief with various arguments for why the court should deny the request. The document included two legal clauses meant to support the police's position, but the defense attorney, after reviewing the brief, argued before the judge that it appeared to have been generated by artificial intelligence, as he could not find any reference to the cited clauses in Israel’s law books.
A police representative was forced to admit that AI had indeed been used. “What was quoted is incorrect. Whoever wrote the document did so mistakenly. An error occurred.”
The judge, in turn, ruled that the police had disgraced themselves. “The law cited does not exist in the State of Israel’s legal code. It hasn’t even existed in anyone’s imagination until now, because searching its wording online via search engines like Google yields no matching results. It was created by artificial intelligence. And if I thought I’d seen everything in my 30 years on the bench, apparently I was mistaken.”
At the start of the hearing, though not in its opening moments, State representatives admitted that the cited law indeed did not exist.
The judge concluded, “In my view, this mistake should never have happened in the first place, and if they were aware of it, they should have informed the court prior to the hearing to avoid further embarrassment.”