Artificial Intelligence (AI) is revolutionizing multiple industries, and the legal world is no exception. With its ability to analyze vast amounts of data and generate insights, the question arises: Can AI write laws? Charles Darwin University (CDU) academic, Associate Professor Guzyal Hill, sought to answer this burning question by testing ChatGPT’s capability to compare, analyze, and produce domestic violence legislation in collaboration with the Australian Law Council.
Domestic violence is a complex and deeply human issue, with up to 50 women losing their lives to it each year in Australia alone. As governments work towards ending violence against women and children, Associate Professor Hill wondered if ChatGPT could contribute to the development of high-quality recommendations and legislation in this area.
However, after conducting several tests and comparing the definition produced by ChatGPT with that of the Australian Law Council, the results showed that human drafting still surpasses AI. While ChatGPT proved useful in classifying and identifying patterns of domestic violence, it fell short in producing a high-quality definition.
Associate Professor Hill, a lawyer and former legislative drafter, cautioned against relying on AI language models (LLMs) like ChatGPT for legal advice, especially for non-lawyers. He emphasized that legal systems differ between countries and even within regions, rendering the application of AI models in legal situations impractical. Notably, ChatGPT now includes a disclaimer acknowledging its inability to provide legal advice.
Recognizing the growing prevalence of AI, Associate Professor Hill stressed the need for lawyers and law students to upskill in this area. Ignoring or eluding AI could lead to unpredictable drawbacks and dangers, including major mistakes in the misuse of AI, missed opportunities to shape the future of law alongside AI, and the risk of experts from other fields developing solutions that neglect human rights or contradict the rule of law.
AI, when used unchecked, poses serious risks and threats. Therefore, lawyers and law students should approach AI practically, cautiously, and curiously. AI systems should augment human acuity rather than replace legal analysis and reasoning. This new AI terrain presents an opportunity for the legal profession to transform and rethink the way law is practiced globally.
In conclusion, while AI shows promise for assisting legal professionals, the research conducted by Associate Professor Hill highlights that human drafting remains superior in producing high-quality legislation. As the legal profession moves forward, it is crucial to embrace AI sensibly and navigate its potential risks and rewards with a keen understanding of the law’s intricacies and human rights considerations.
