ABSTRACT The article examines the potential implications of ChatGPT ChatGPT (Generative Pre-trained Transformer) [Alec Radford and others, ‘Improving Language Understanding by Generative Pre-Training’ Amazon Web Services (2018) <https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/language-unsupervised/language_understanding_paper.pdf> accessed 28 May 2024; When this article was submitted, only ChatGPT-3.5 and ChatGPT-4 was available. However, during the revision stage, ChatGPT-4o (also known as GPT-4 Omnus) was introduced to the market, as announced on 13 May 2024.] in judicial decision-making process. This is especially salient given that judicial decisions are made within an interactive ritual chain (Bergman Blix) rather than an ivory tower. This cutting-edge natural language processing model leverages deep learning techniques to potentially aid judiciary members in formulating legal judgements. Therefore, this article assesses the capabilities, limitations, and potential applications of ChatGPT, aiming to evaluate the model's feasibility as a collaborative contributor to legal judgements. In particular, the article examines the utilisation of ChatGPT as an adjunctive tool, supporting human judges in their decision-making processes and, consequently, establishing to what extent artificial intelligence (AI) is used as a tool or collaborator in judicial decision-making. This determines authorship and, accordingly, ownership of judgements.
Read full abstract