The Brazilian city of Porto Alegre recently introduced an ordinance that was later revealed to have been written by ChatGPT, an artificial intelligence language model. The municipal council member, Alderman Ramiro Rosario, intentionally concealed this fact when submitting the proposal. The ordinance aimed to eliminate the requirement for taxpayers to pay for stolen water meters’ replacements—a recurring issue in the city.
Rosario enlisted the assistance of ChatGPT by providing a short 49-word prompt outlining the objective. Without disclosing the involvement of artificial intelligence, he presented the proposal to his 35 council colleagues. Rosario justified his decision by stating, If I had revealed it earlier, it would not have even been possible to vote on the proposal. It would be unfair to the city’s residents not to approve the project just because it was written with artificial intelligence. Subsequently, the proposal received unanimous approval, and the decree went into effect on November 23.
Emphasizing the potential of technology, Rosario expressed his conviction that this marked a new technological revolution with the ability to utilize tools for both good and evil. He stated, That’s why we need to show how they can be used for good.
The revelation of AI involvement occurred when Council President Hamilton Sosmer learned about Rosario’s use of ChatGPT through a social media post. Initially viewing it as a dangerous precedent, Sosmer later changed his stance, recognizing that it was likely to become a common practice. I started looking into it more and saw that, for better or worse, this was going to be a trend, he stated.
The use of ChatGPT in legislative drafting is not an isolated incident. In one notable example, Massachusetts Democratic Senator Barry Feingold employed the chatbot to assist in drafting a proposal aimed at regulating AI. Senator Feingold believes that chatbots can be valuable tools in addressing the complexities of creating legislation, particularly in consulting existing laws.
Transparency emerged as a crucial aspect of using artificial intelligence in these contexts. Senator Feingold stressed the necessity of watermarking work created with ChatGPT and ensuring clarity regarding its involvement. I support people using ChatGPT to write proposals, as long as they are clear, he affirmed during an interview.
The advent of technology in the legislative process marks a significant shift, inviting discussions on its potential benefits and challenges. As the world continues to evolve, embracing AI in innovative ways, instances like these demonstrate the need for transparency and ongoing exploration of ethical implications.
With artificial intelligence increasingly playing a role in various fields, the ethical utilization of these technologies remains a critical consideration. The example from Porto Alegre, while controversial, highlights the power of AI in shaping legislation and presents an opportunity to examine the appropriate boundaries and ethical guidelines necessary moving forward.
