Exploring LLMs Applications in Law: A Literature Review on Current Legal NLP Approaches

The integration of Natural Language Processing (NLP) and AI into legal tasks is a natural progression, given the linguistic nature of law. This combination allows for more efficient and accurate analysis of legal texts, enhancing various aspects of legal practice. The use of NLP and AI in legal tech has a long history, dating back to the 1960s with the development of online legal content search systems [4], [5], [6]. NLP, particularly with the advent of Large Language Models (LLMs), has made significant strides in legal applications, aiding in tasks requiring language processing and understanding. LLMs represent cutting-edge technology, advancing AI approaches in various domains involving textual contents such as medicine [7], [8], [9], engineering [10], [11] and law [12], [13], [14], [15]. Overall, the integration of AI and NLP in the legal domain holds great promise for improving efficiency and decision-making processes across various legal tasks. However, addressing challenges related to context, data availability, and interpretability remains essential for the reliable application of these technologies in the legal domain [16], [17]. Within this context, applications refer to the broader use cases where AI technologies address specific legal challenges or fulfil specific needs, such as contract review, document automation, and compliance monitoring.

2. CASE ENTAILMENT (T2)

Case Entailment is the task of determining whether the facts, legal principles and arguments introduced in a legal case logically support or imply the outcome of another case, requiring to analyse the content of legal documents to identify relationships and inferring conclusions based on precedents [18], [44]. We have found 4 studies that focus on this task (P06, P07, P13, P28), and their main legal area of the application is Case Law.

In P06 [59], out of 4 tasks, Task 2 (a case law entailment task) aims to distinguish which paragraph in a supporting case implies the provided text fragment and Task 4 (statutory entailment task) focuses on determining the potential implication of a bar exam question by a set of relevant articles. For Task 2, the jurisprudence implication task requires finding an implying paragraph from a case,

3. QUESTION ANSWERING (T3)

This task is focused on the application of NLP and machine learning techniques to automatically provide answers to legal queries, which involves understanding and interpreting complex legal questions and retrieving or generating accurate and relevant answers based on legal texts such as statutes, case law, regulations, and legal opinions [18], [45]. We have found two studies for this task (P49 and P53) which are embedded within the Legal Research and Miscellaneous legal areas.

Understanding legal texts presents a considerable challenge due to their extensive and intricate clauses, compounded by a scarcity of datasets annotated by experts.

FRAQUE operates by querying unstructured data from various sources such as documents, websites, and social media. Leveraging statistical components like word embeddings, the system offers flexibility in adapting to different domains and languages. The primary objective of FRAQUE is to match questions with relevant frames and corresponding document passages stored in a knowledge graph, which are then presented as answers. To ensure user-friendliness, FRAQUE’s development follows a usercentered design approach, allowing for the monitoring of linguistic patterns used by users and identifying the most frequently occurring structures in their queries. Both P49 [76] and P53 [77] highlight the importance

In the context of legal AI, this tasks focus on measuring how closely related or alike two pieces of legal text are

Domain pretraining with a custom legal vocabulary exhibited the most substantial performance gains,

these studies highlight the potential of text similarity, document clustering, and topic modelling techniques for legal document analysis. The proposed methods have shown promising results in improving the performance of legal document analysis tasks, such as law article retrieval, case retrieval, and topic modelling. However, further research is needed to address the challenges posed by the complexity and variability of legal language and the need for large annotated corpora. The development of more sophisticated techniques for legal document analysis could have significant implications for the legal industry, enabling more efficient and accurate analysis and retrieval of legal documents.