Context
The advent of generative artificial intelligence is radically transforming the legal sector, reshaping the way lawyers, judges, and public administrations search for, analyse, and draft legal documents.
While this technology offers extraordinary efficiency opportunities, it also introduces critical and unprecedented risks for the administration of justice: the generation of plausible but false content (hallucinations), algorithmic bias, privacy issues, and the risk of deskilling the profession. In this scenario, it becomes essential to develop tools to govern innovation rather than be driven by it.
Purpose
The main objective of GenAI4Law is to guide the understanding and adoption of GenAI in the legal domain in a responsible, safe, and transparent manner. The project integrates computer science, law, and ethics through three main research lines:
- Line 1 – Analysis of the state of the art and risks: Analyse the current state of GenAI in the legal domain, map development and deployment scenarios, and critically assess risks and opportunities for the legal community.
- Line 2 – Methods and tools for reliable legal applications: Develop methods for the effective use of GenAI in law, including innovative models and prototypes capable of providing legally meaningful explanations and ensuring controllability of outputs and reasoning.
- Line 3 – Evaluation, criteria, and ethical-legal benchmarking: Define technically rigorous criteria for evaluating and comparing GenAI applications in law, ensuring their use is lawful, ethical, and aligned with the values of the legal system.
Timeline and Structure
Total duration: 54 months
Leadership and Team: Led by Principal Investigator Prof. Giovanni Sartor, GenAI4Law is based at the University of Bologna. The activities integrate cross-disciplinary expertise in legal theory, artificial intelligence, and legal informatics.
