In a recent twist to legislative processes, a city in Brazil unwittingly became the stage for what could be the nation’s first AI-crafted law, entirely generated by artificial intelligence. Councillor Ramiro Rosario, without the knowledge of his peers, employed OpenAI’s chatbot, ChatGPT, to draft a proposal addressing the replacement of stolen water consumption meters. The subsequent unanimous approval by the council has sparked a debate on the role of AI in shaping public policy, revealing both the potential and pitfalls of this emerging technology.
In October, Porto Alegre, the second-largest city in Brazil’s south, witnessed the silent passage of a law that, unbeknownst to the councillors, was authored entirely by AI. Councillor Ramiro Rosario, seeking a solution to the issue of stolen water consumption meters, turned to ChatGPT to generate a proposal. The unsuspecting council, consisting of 36 members, approved the bill unanimously, with the law officially coming into effect on November 23.
The arrival of ChatGPT in the public sphere a year ago initiated a global conversation on the transformative potential of AI-powered chatbots. While some see it as a promising tool to streamline legislative processes, concerns linger about the unintended consequences of machines handling tasks traditionally executed by humans. The incident in Porto Alegre serves as a stark example of the dual nature of AI in public policy, raising questions about transparency and accountability.
AI, driven by large language models like ChatGPT, operates by predicting the next word in a sentence. Yet, this process is not foolproof, as these models may introduce false information—a phenomenon known as hallucination. Research by tech company Vectara suggests that even advanced models, such as GPT, can generate false information up to 3% of the time, raising concerns about the reliability of AI-generated legislative content.
Legal implications of AI-crafted law
The use of AI, particularly in legal contexts, has sparked discussions about its ability to comprehend the nuances and complexities inherent in the law. Dean Andrew Perlman of Suffolk University Law School warns that while ChatGPT may signify a monumental shift, its limitations become evident in situations requiring in-depth legal analysis. The potential shortcomings of AI in interpreting legal principles and precedent raise concerns about its role in shaping the legal landscape.
Despite the lack of transparency surrounding the AI-crafted proposal, Councillor Rosario defends his intentional secrecy. He asserts that revealing the AI origin beforehand might have jeopardized the project’s approval. Rosario claims to have entered a concise prompt into ChatGPT, receiving a full draft proposal within seconds, complete with justifications. His motive extends beyond resolving a local issue, aiming to provoke a broader debate on the ethical use of technology in governance.
Initially labeling the AI-authored legislation as a “dangerous precedent,” Council President Hamilton Sossmeier underwent a change of perspective. Upon delving into the matter more deeply, he acknowledges the inevitability of AI influencing legislative processes, dubbing it a “trend.” The incident in Porto Alegre prompts reflection on the evolving relationship between AI and governance, with the potential for a paradigm shift in legislative practices.
The unseen influence of AI in legislation
As Porto Alegre grapples with the aftermath of an AI-crafted law slipping through the council unnoticed, the broader implications of AI in legislation demand scrutiny. The incident raises fundamental questions about transparency, accountability, and the ethical use of technology in shaping public policy. As society navigates this uncharted territory, one must ponder: Are we prepared for a future where AI plays a pivotal role in crafting our laws, and how do we strike a balance between innovation and safeguarding the integrity of our democratic processes?
