AI4LEGS- ECAI2025: Legal-informatics approaches to LLMs & Law in legislation

Co-located with the 28th European Conference on Artificial Intelligence, which will take place on October 25-30, 2025 (ECAI2025)

Objectives

This tutorial explores the transformative role of Artificial Intelligence (AI) in the legal sector, focusing on how AI tools enhance legal research, practice, and decision-making, with particular focus on Large Language Models (LLMs). It aims to:

• Introduce an expert in AI but non-specialist in a legal domain to the legal methodology and evaluation.
• Introduce an expert in the legal domain to the most important foundations of AI, with particular regard to LLM.
• Motivate and explain a topic of emerging importance for AI in institutions, with particular regard to the explicability
• Present some use-cases in Parliaments or the Assembly with invited speakers.

Audience

The primary audience includes researchers - PhD and experienced researchers also in R&D departments of the companies - in interdisciplinary areas with a foundational understanding of AI or legal disciplines. Participants are expected to have basic familiarity with legal processes and a general awareness of AI applications.

Knowledge acquired and methodology

Attendees will gain insights into AI techniques in the legal domain, such as NLP, Machine Learning, and LLM (60%). The tutorial will also address ethical and regulatory considerations (30%) pertinent to deploying AI in legal settings. Finally, the tutorial would also like to address the topic of explicability and the necessity to have a clear, transparent, complete visualization of the results using human-computer interaction techniques.

The session will cover AI applied to legal information, development and deployment of AI models for tasks like legal information retrieval, legal drafting, compliance checking, legal information extraction for policy makers, visualization of the results for XAI, and so forth.

Practical demonstrations will showcase state-of-the-art AI-driven tools for legal research, document drafting, and support systems in legislative drafting or in the policy making stage. These systems will be presented alongside limitations and opportunities for the use of AI and LLMs in the legal domain, especially in light of the AI Act and the Ethical principles. Live demos and experiments will be conducted alongside theoretical reflexions

Innovation

Integrating AI into the legal field is revolutionizing traditional legal practices, offering increased efficiency and accuracy. However, introducing AI in Parliaments, Assemblies, and Governments requires special attention to the theory of law constraints, democratic principles, and ethical issues. These elements should be included by-design in the modeling of AI, in the methodology of the AI solutions, and in the lifecycle of the applications maintenance (human-in-the-loop). Understanding these advancements is crucial for legal professionals to remain competitive and ethically informed, but also for computer scientists and engineers to correctly implement the solutions to support the institutions. At the same time, the new regulatory landscape (AI Act) will be partly explored.

Proposers

Monica Palmirani, University of Bologna, ALMA AI

Monica Palmirani is full professor in Computer Science and Law and Legal Informatics at University of Bologna, School of Law. She co-chairs the LegalDocML and the LegalRuleML. Since 2013 she serves on the OASIS LegalXML Steering Committee. In 2015, she was recognized as an OASIS Distinguished Contributor. She was member of Board of Directors of OASIS from 2016 till 2018. Her research fields include XML techniques for modelling legal documents in structure as well as in aspects connected to legal knowledge, including logic rules and legal ontologies, and ICT-enhanced legal drafting techniques using artificial intelligence techniques. She is also the scientific coordinator of the Legal Blockchain Lab and teaching course in Law School about DLT. She has published more than 130 papers and she has been chair of several international conferences, editor of book series and member of the scientific committee of “AI and Law” Journal. She is president of the IAAIL (International Association for Artificial Intelligence and Law), pointed out as President for the period 2024-2025. She is Director of the International PhD programme “Law, Science and Technologies” MSCA-ITN. She has winner of the ERC Advanced Grant of the European Research Council for five years of ground-breaking topics with a funding of 2.5Ml of Euro. She is also principal investigator of Jean Monnet Module on Legal Design.

Other contributors:

• Salvatore Sapienza, University of Bologna, ALMA AI
• Michele Corazza, University of Bologna, ALMA AI
• Fabio Vitali, University of Bologna, ALMA AI
• Camera dei Deputati
• Regione Emilia-Romagna
• European Publication Office
• European Commission

Programme

Detailed outline of the tutorial:

1)       Foundations of AI in Legal Practice (30 minutes)

• Introduction to the peculiarities of the legal domain and the legal language and its relationship with AI. The Hybrid AI methodology.
• The Akoma Ntoso (AKN) standard: the relevance of machine-readable, high-quality legal data
• Overview of AI technologies applied to the legal domain: how AI applications transform legal research, practice, and decision-making

2)      AI Techniques and Tools for Legislative Domain: opportunities and limitations (30 minutes)

• NLP in Legal Contexts: exploring the limits of NLP methods in analyzing legal documents and extracting pertinent information, discovering hidden information, mapping policies with the legislation
• ML applications in Legal Contexts: exploring the representation of legal texts in AI systems in probabilistic models
• Introduction to Large Language Models (LLMs) in the legal domain: understanding the role of LLMs in processing and generating legal information

3)        XAI in legal domain and visualizaiton techniques  (30 minutes)

• HCI and XAI visualization techniques: exploring some techniques for XAI in legal domain
• Legal design methodology for presenting AI results and guarantee the autonomy of the decision-makers

4)        DEMOs: Practical applications of AI in legal tasks (30 minutes)

• Legal Information and AKN:  demonstrations of AI-driven experiments leveraging on semantic technologies
• Automated legal drafting: demostrations of AI-assisted drafting for legislative provisions
• Chamber of Deputy of Italy use-case
• Regione Emilia-Romagna use-case: SAVIA project

5)       Ethical and Regulatory Considerations (30 minutes)

• Ethical implications: discussing bias, transparency, and accountability in AI applications for legislation
• Legal implications: Understanding the EU AI Act, GPAIs, systemic risks, and their impact on legislation.
• Technical mitigation measures and possible approaches to AI modelization.

6)      Future Trends and Challenges (30 mins)

• What solutions? Exploring advancements (Hybrid and Neurosymbolic AI) in AI and their potential applications in law.