Illustration by Tag Hartman-Simkins/Futurism. Source: Getty Images
The point of an apology is that it must be sincere. It also helps if you put some real effort into making it.
A judge in New Zealand is being forced to consider these criteria after receiving a written apology from a defendant, which appears to have been generated by – and why we’ll be talking about it on this website – an AI model.
The defendant had a lot to answer for: He pleaded guilty to arson and other charges after burning down a house. While in custody he also bit a police officer, after which he apparently “felt pleasure” in telling the police that he had AIDS, new zealand herald first reported.
When it came time for sentencing, Judge Tom Gilbert of the District Court in Christchurch asked the defendant about apology letters he had sent, one to the judge and one to the victims.
“The issue of repentance is interesting,” Gilbert said, according to a transcript. Viewed by new York Times. He continued, “Out of curiosity I used two AI tools to ‘draft a letter for me to a judge, in which I expressed remorse for my crime.'” “It immediately became clear that these were two AI-generated letters, albeit with changes around the edges.”
Remorse is a factor in sentencing. If you truly feel sorry for your actions, you may receive a lesser sentence. And the judge was not against the use of AI, he clarified.
Gilbert said, “But certainly when one is considering the genuineness of a person’s remorse, as far as I’m concerned, simply producing a computer-generated letter doesn’t really get me anywhere.”
Ultimately, he decided to reduce his sentence by five percent, while his lawyers were demanding a ten percent reduction. She will now spend 27 months in jail. “I do not believe this is a case where 10 percent is appropriate and, in fact, 5 percent could be considered appropriately generous,” Gilbert said.
The role of AI in the courtroom remains a source of confusion and controversy. Lawyers have been warned by judges after being caught using an AI tool that left hallucinatory passages in official filings, including faulty citations and fabricated case law. These types of incidents have created mini-crises in many law firms.
This isn’t even the first apology-related AI courtroom mistake. Last October, after being caught using hallucination AI in court filings, a defense attorney submitted a brief explaining his AI use. Too Written with AI – something they first denied, later apologized for, and then went back to admitting using. If there’s a lesson to be learned here, it’s that judges have long been wise to the proliferation of lazy AI technology.
More on AI: Law firm fires hundreds of employees amid turn to AI
