In what could be a precedent for the use of AI, a law drafted by ChatGPT was passed by the Porto Alegre council (a city in Brazil). The law prohibits the government from charging citizens for replacing water meters that were stolen by thieves.
However, none of the 36 councilmen who unanimously passed the bill had any idea of the use of AI. Councilman Ramiro Rosário later revealed that he had taken the help of ChatGPT to design the proposal.
To everyone’s relief, it wasn’t the AI chatbot that birthed the idea of the new law. Rosário used a 49-word prompt to direct ChatGPT to write the proposal.
He admitted that if he had revealed the use of AI in the process, the council wouldn’t have taken up the matter in the first place.
Hamilton Sossmeier, the city council president, initially didn’t approve of Rosário’s methods, believing that it was a “dangerous precedent.”
However, he later remarked, “I started to read more in depth and saw that, unfortunately, or fortunately, this is going to be a trend.”
The Dangers of AI
This isn’t the first time AI has come under the scanner in the legal field. Earlier this year, two New York attorneys, Steven Schwartz, and Peter LoDuca, came under fire for using ChatGPT for citing non-existing cases.
They used the AI chatbot to find cases similar to the one they were fighting. To their delight, ChatGPT came up with legitimate-looking answers.
However, when cross-examined, the cases were found to be fake. ChatGPT had created bogus cases, which never really took place. The lawyers had to pay a penalty of $5,000, and the case was dismissed.
This incident sheds light on the unreliability of AI chatbots. According to Andrew Perlman, dean at Suffolk University Law School, AI chatbots aren’t a replacement for human lawyers. It lacks understanding, judgment and the skill of interpretation of the legal framework.
In late May 2023, Judge Brantley Starr, a Texas federal judge, made it compulsory for every lawyer presenting a case before him to file a declaration that no part of the submissions was drafted with the help of AI.
He had to resort to this “Mandatory Certification Regarding Generative Artificial Intelligence”, after a lawyer (Steven Schwartz) was found using AI writing tools for six of his cases.
Another bill, drafted by ChatGPT, is yet to be voted on in Massachusetts. Sen. Barry Finegold has drafted this bill with the help of the AI Chatbot, which aims to regulate the use of artificial intelligence. He also advocates the watermarking of content generated by ChatGPT.
With AI taking center stage in many professions and businesses, it remains to be seen how the legal realm reacts to the use of AI.