In a surprising turn of events, lawmakers in a Brazilian city inadvertently approved a piece of legislation authored by ChatGPT, an artificial intelligence program. The legislation, which was enacted in the city of Porto Alegre in October, was revealed this week to have been written using the chatbot by city council member Ramiro Rosário.
According to the Associated Press, Rosário had requested ChatGPT to draft a proposal aimed at preventing taxpayers from shouldering the cost of replacing water consumption meters. He then presented the proposal to his fellow city council members without disclosing its origin, fearing that revealing the involvement of artificial intelligence would hinder its chances of approval.
“It would be unfair to the population to run the risk of the project not being approved simply because it was written by artificial intelligence,” Rosário explained to the AP. He believed that disclosing the chatbot’s involvement would have prejudiced the council’s decision-making process.
Interestingly, this is not the first instance of ChatGPT being involved in the drafting of legislation. In the United States, Massachusetts State Sen. Barry Finegold enlisted the help of ChatGPT to assist in the creation of a bill aimed at requiring companies using AI chatbots to implement risk assessments and security measures to enhance government understanding of their algorithms.
The use of ChatGPT in legislative drafting raises important questions about the role of artificial intelligence in governance, prompting the urgent need for government regulations on the use of chatbots in official decision-making processes.
