The council of Porto Alegre, a city in southern Brazil, has recently passed legislation written by an AI chatbot, ChatGPT.
The new ordinance is aimed at preventing the city from charging taxpayers to replace stolen water meters. The proposal received unanimous approval from 36 members of the council and took effect in late November. Surprisingly, it was only after the passing of the proposal that councilman Ramiro Rosário revealed that the text was generated by an AI chatbot called ChatGPT.
Rosário admitted to using a 49-word prompt to instruct OpenAI’s chatbot to generate the complete draft of the proposal. There were initial concerns surrounding the use of AI to write legal documents, but some council members, like Hamilton Sossmeier, expressed a change of heart after realizing that such methods may become a trend in the future.
He may be correct, as similar initiatives have taken place in the US, with lawmakers drafting a bill using AI to regulate generative artificial intelligence models like ChatGPT. This new approach to legislation creation, however, raises controversial and risky consequences. Attorneys in New York who referenced ChatGPT in a lawsuit faced complications when the AI produced fabricated information.
Despite the risks, advocates like Rosário believe that such technology can be used effectively for good. However, it’s crucial to recognize the limitations and potential consequences, as seen with legal cases that cited the AI’s generated but fictitious information. Awareness of these limitations is essential, as demonstrated by the fine and dismissal of the lawsuit in the New York case.
As the use of AI continues to evolve, the technology presents numerous possibilities but also demands careful consideration and understanding of its capabilities and limitations.
