City lawmakers in Brazil have approved an ordinance that, unbeknownst to them, was drafted by artificial intelligence—sparking further debate about the place of AI in lawmaking.
Ramiro Rosário, a councilman in the southern city of Porto Alegre, bragged on X, formerly Twitter, last week that he had concocted a proposal with the help of ChatGPT.
Rosário said that the use of AI to craft the ordinance, which aims to prevent government officials in the city of 1.3 million from charging residents to replace stolen water meters, would have gone unnoticed had the councilman not flagged the public after the fact.
“There is only one lesson to be learned: technology serves to reduce costs and optimize our work. It will bring gains in quality and productivity,” Rosário posted via X on Nov. 29.
“No one will be replaced by AI, but by those who know how to use it,” Rosário said.
In an interview, Rosário told Law.com International that his experiment has shed more light on AI’s possibilities but acknowledged that not everyone agreed with him.
“This experiment in Porto Alegre has opened a worldwide debate on the use of technology to enhance quality, increase productivity and reduce costs in public administration in Brazil and across the world,” he said. “I support the idea that artificial intelligence can help optimize resources and the time of political agents and public servants, allowing them to focus on what is truly essential for their work.”
Reactions have poured in from around the globe.
Edge Pereira, a Brisbane, Australia-based associate director at KPMG who focuses on enterprise architecture and information technology strategy, said via a LinkedIn post that Rosário’s use of AI to draft legislation “raises significant ethical concerns.”
“Firstly, there’s the issue of transparency and informed consent,” wrote Pereira, whose profile lists academic credentials that include a bachelor’s degree in computer science from the Universidade de Fortaleza in Brazil, as well as studies on artificial intelligence in business strategy at MIT’s Sloan School of Management.
“The revelation that AI, rather than a human, authored the law only after its passage undermines the democratic process, as stakeholders were not fully informed,” Pereira wrote.
“Secondly, this scenario questions the accountability and responsibility in law-making. If an AI writes a law, determining liability for any unintended consequences or flaws in the legislation becomes complex,” Pereira added.
“Thirdly, there’s the potential for AI to embed biases in the law, reflecting the data it was trained on, which may perpetuate existing inequalities,” Pereira wrote.
“Lastly, this incident challenges the traditional understanding of expertise and authority in legislative processes, potentially diminishing the role of human judgment and expertise in crafting laws tailored to societal needs and values,” Pereira added.
Law.com International has reached out to Pereira for additional comment.
Unapologetic
Rosário has been unapologetic, telling Law.com International Monday that he didn’t disclose his use of ChatGPT ahead of the vote so as to guarantee approval of the proposal while also validating “the beneficial use of technology.”
Many Brazilian lawmakers “are prejudiced against the use of technology and are uninformed about the discussions happening globally on this subject,” Rosário said via a private message from his X account.
“We are on the brink of experiencing a technological revolution even more incredible and impactful than the use of computers and the advent of the internet,” he added.
The experiment in Porto Alegre, Rosário hopes, will show Brazil’s more than 5,000 municipalities and city councils that AI can contribute to efficiencies within the country’s extremely complex and vast legal framework.
By leaning more on machines, he argues, the country can reduce administrative costs.