RESEARCH PROJECT TITLE: Element-Aware Legal Aid AI System
The rapid development of artificial intelligence (AI) technologies has profoundly impacted the legal industry, driving innovation in areas such as legal document automation, case retrieval, and legal reasoning. Furthermore, many existing legal AI systems produce outputs in a black-box manner, lacking transparency and interpretability, which undermines trust among legal professionals. Also, a key issue lies in the reliance on manually predefined legal elements, a process that is time-consuming, incomplete, and often non-transferable across legal contexts. Furthermore, the availability of high-quality legal ontologies and datasets, and their integration into user-centric systems remains limited, and existing systems often neglect user experience design, making them inaccessible to vulnerable groups and individuals with limited legal knowledge. To address these challenges, this paper presents the Element-Aware Legal AI System, which combines data-driven and knowledge-driven approaches to enhance logical consistency, interpretability, and trustworthiness in legal tasks. We introduce a generalized legal element definition scheme to improve structural logic, comprehensiveness, and adaptability, leveraging supervised, unsupervised, and self-supervised learning methods, as well as large language models (LLMs), to automatically construct a precise legal knowledge graph representing legal elements and their relationships. To bridge the accessibility gap, we incorporate legal elements into intuitive page design and interactive workflows, ensuring user-friendliness and practical usability. By doing so, the system empowers underserved communities and individuals who cannot afford professional legal services, providing much-needed legal assistance and fostering equitable access to justice.
Supervisors: Antonino Rotolo, UNIBO
Institutions involved in the co-tutelle: in progress
https://www.unibo.it/sitoweb/qingjing.chen2/