A woman in New Zealand has been sentenced to over three years in prison for burning down a house she rented and assaulting a police officer. Normally that wouldn’t be the kind of thing that makes international news. Except last week the judge in the case noticed something about the 37-year-old defendant’s apology letter that’s likely to be a consideration for all kinds of professions well into the future. The woman apparently used generative AI tools to write her apology. And the judge called into question whether that meant she was really sorry.
Michae Ngaire Win was being sentenced last week for arson, burglary, assault, and resisting police, according to a report from the New Zealand Herald, when Judge Tom Gilbert became “unimpressed” by the apology letters she had produced to the owner of the home she burned down as well as the first responders she assaulted. Win bit a cop and reportedly “took delight” in telling the officer she had AIDS.
“Out of curiosity I punched into two A.I. tools ‘draft me a letter for a judge expressing remorse for my offending,” the judge reportedly said, according to a transcript read by the New York Times. “It became immediately apparent that these were two A.I.-generated letters, albeit with tweaks around the edges.”
The judge reportedly seemed open to the idea that using AI wasn’t necessarily a bad thing. But he made the point that “when one is considering the genuineness of an individual’s remorse, simply producing a computer-generated letter does not really take me anywhere as far as I am concerned.”
The question, of course, is whether a human can ever take ownership of something produce by an AI chatbot. Some people argue that you can, especially if you’re prompting the robot with plenty of specifics about what you’re looking for. Some make the comparison to photography, where humans are getting substantial assistance from a machine to create an image, contrasting it with the more “pure” art of drawing or painting.
But plenty of other people will argue that you’re not actually producing something that would allow you to take ownership with mere prompts. The U.S. Copyright Office agrees with folks in this camp, given the fact that AI-generated works can’t be registered under American copyright law.
Obviously the world is still trying to figure out how to properly assess the validity of AI-generated work. Kids are constantly using tools like ChatGPT to do their homework, raising questions about what literacy might look like in the near future.
Generative AI’s ubiquity makes it feel like the genie is out of the bottle. But we still have quite a ways to go until we figure out the social norms around its use. In the meantime, it’s probably a good idea not to rely on consumer-available AI too heavily for things that matter, like writing legal documents or dispensing medication to a loved one.
Court apologies likely seemed like a safe use to this woman in New Zealand, but you also have to consider what happens if you get caught. Because “sorry” doesn’t sound quite the same if the judge knows you didn’t write it yourself. A pre-sentencing report had suggested that the woman serve no jail time for her arson and instead be confined to home detention, according to the New Zealand Herald. The judge handed down a sentence of 27 months in prison.
Read the full article here
