An analysis of the current limitation and future directions of AI applied to the legal domain based on a SLR and early prototyping

The objective of the thesis is to conduct an analysis of the current state of artificial intelligence (AI) applied to the legal domain, examining its limitations and improvements over the years. Furthermore, attention will be dedicated to the most promising directions for future enhancements. At the end, we will present a prototype, a proof of concept, for handling the process of resolution between conflicting laws. As highlighted in multiple papers, the legal domain poses unique challenges for AI applications. It demands advanced reasoning capabilities, the ability to interpret complex language structures, and precise decision-making based on legal precedents and contextual understanding. The nature of legal texts demands robust and adaptable models capable of capturing both semantic nuances and logical structures. Working in the legal domain could drive advances in deep learning techniques due to the difficulty faced and be a stimulating challenge for researchers. In particular, we will focus on the topics of legal reasoning and information retrieval. Regarding the latter, we will also consider its subdomains, such as information extraction and legal information retrieval.