Legal | Language Technologies Lab

Argument Mining on Legal Datasets

Mon, 02 Mar 2026 00:00:00 +0000

Description:
Argumentation in legal documents is typically well-structured and follows specific domain rules. We are interested in applying both the most recent NLP techniques and also hybrid techniques that can leverage the domain knowledge.

Contact: Andrea Galassi, Marco Lippi

Transformers and LLMs for the detection and classification of unfair clauses

Mon, 02 Mar 2026 00:00:00 +0000

Description:
For several years, we have been working on tools for the automatic detection of unfair clauses in Terms of Services and Privacy Policies documents in the English language (see CLAUDETTE and PRIMA Projects page). We have already conducted several studies on this topic, and we are interested in applying new effective methods and techniques. Right now, we are focused on LLMs, but we are also interested in alternative techniques.

Contact: Andrea Galassi, Marco Lippi

AMELR

Fri, 27 Feb 2026 00:00:00 +0000

The AMELR workshop focuses on Legal Argument Mining (LAM) - using NLP to automatically detect legal arguments. Recent developments in NLP and LAM have provided legal scholars with a powerful tool for studying reasoning patterns, interpretative theories, and biases across jurisdictions and legal systems. The workshop gathers experts in computer science, AI & Law, legal theory, and empirical legal studies to address key challenges of LAM: creating training datasets, developing reliable models, establishing reproducibility standards, and integrating LAM into legal research. The workshop aims to strengthen the emerging field of LAM and its role in empirical legal studies by sharing latest implementations, addressing core challenges, and establishing best practices.

ASAIL

Fri, 27 Feb 2026 00:00:00 +0000

The ASAIL workshop series and interest group serves as a platform for researchers and practitioners working on natural language processing of legal text. Its goals include (i) Organising regular peer-reviewed workshop events for presentation and discussion of research and practical implementations around legal NLP; (ii) Facilitating communication and collaboration among academic researchers as well as practitioners from industry, government, and the public sector, and other interested individuals and organisations; (iii) Providing an entry point into the research field and community.

Legal Analytics

Fri, 27 Feb 2026 00:00:00 +0000

Argumentation and Argument Schema

We have defined legal arguments and developed benchmarks for training machine learning in automatically identifying them. Legal arguments are relevant to several legal tasks (e.g., judgement prediction) as they encode important decisions and opinions that are tailored to a given problem. Moreover, legal experts are interested in assessing if there are some recurring patterns concerning legal arguments. For instance, do similar documents convey similar legal arguments? Can we compare documents based on their arguments?

Judgement Prediction

The automatic prediction of the judge’s decision. We have developed benchmarks and assessed standard and transformer-based models on this task.

Interpretability

We have explored memory-augmented neural networks and decision trees to define more interpretable models, encouraging user trustworthiness based on explanations in addition to model efficiency.

Unfair Clause Detection

The automatic identification and classification of unfair clauses. We have developed benchmarks on this topic.

Summarization

Can machine learning models summarize legal documents based on certain guidelines? Are the developed summaries useful to legal experts?

Cross-linguality

We have developed some methods to project labels from similar documents written in different languages.

Information Retrieval

The automatic retrieval of legal documents and knowledge based on a similarity metric. This also link to the argumentation topic if we use arguments (either quantitatively or qualitatively) to compare and rank retrieved documents.

EquAl: Equitable Algorithms, Promoting Fairness and Countering Algorithmic Discrimination Through Norms and Technologies - Final Conference

Fri, 23 Jan 2026 00:00:00 +0000

Project info

Useful Links

PRIMA: PRivacy Infringements Machine-Advice - Final Conference

Mon, 12 Jan 2026 00:00:00 +0000

Project info

PRIMA (PRivacy Infringements Machine-Advice) studies the law and practice of privacy policies, develops methods and techniques for their automated analysis, and implements a prototype to assess their lawfulness. It deploys legal analytics—a mix of data science, artificial intelligence, machine learning, natural language processing and statistics—to detect and assess privacy policies’ infringements.

Useful Links

Is It Worth Using LLMs for Unfair Clause Detection in Terms of Service?

Wed, 01 Jan 2025 00:00:00 +0000

🏆 Awarded the “Peter Jackson” Award for Best Innovative Application Paper

Principles Of Law In National and European VAT (POLINE)

Wed, 01 Jan 2025 00:00:00 +0000

POLINE aims at developing an AI-powered pilot tool for the retrieval and analysis of judicial principles of law in the CJEU and national case-law in Value Added Tax (VAT). The tool relies on AI techniques for extracting, clustering and linking judicial principles of law and is embedded in a modular platform, consisting of a Legal Database, Link Visualization and Customised Detection Module. It covers the case-law of the CJEU and the Italian, Swedish and Bulgarian Supreme Courts and will be accessible to judges, other legal practitioners, tax policymakers and taxpayers. The development of the tool will be based on a multidisciplinary approach combining theory and practice of judicial decision-making for the study of the concept of “judicial principle of law” and the analysis of the case-law; legal informatics methods for the creation of an ontology of judicial concepts in VAT and training datasets of annotated judicial principles of law; AI, machine learning, and NLP techniques for the automatic extraction of principles, the detection of textual and semantic similarity, and network analysis. The tool will be trialled in 3 online national testing events and disseminated in 3 national demonstration events and 1 final international conference. The pilot tool provides a robust and trustworthy use case of AI technologies for justice. It will provide non-discriminatory and effective access to justice. Through its collection of principles of law and NLP-powered search engine, the tool will assist judges in accessing legal knowledge reducing their work overload. Moreover, through the Customised Detection Test Module, the tool will allow recipients of VAT measures to identify judicial principles of law applied in a specific case and assess whether VAT law is correctly applied. By developing open-access automated techniques of knowledge extraction, the methods developed can be easily reused and expanded to include other fields of law and other legal systems.

PRivacy Infringements Machine-Advice (PRIMA)

Mon, 01 Jan 2024 00:00:00 +0000

TeamUnibo at SemEval-2023 Task 6: A transformer based approach to Rhetorical Roles prediction and NER in Legal Texts

Sun, 01 Jan 2023 00:00:00 +0000

Combining WordNet and Word Embeddings in Data Augmentation for Legal Texts

Thu, 01 Dec 2022 00:00:00 +0000

Add the full text or supplementary notes for the publication here using Markdown formatting.

Equitable Algorithms, Promoting Fairness and Countering Algorithmic Discrimination Through Norms and Technologies (EquAl)

Sat, 01 Jan 2022 00:00:00 +0000

The EquAl project addresses algorithmic evaluations, decisions, and predictions, to promote fairness and counter discrimination affecting individuals and groups. The research project fundedis by the EU Commission under the NextGenerationEU program and the Italian Ministry of Education, University and Research. (PRIN 2022. Ref. prot. n.: 2022KFLF3E-001 - CUP J53D23005560001) EquAl aims (i) to provide an understanding of the concepts of algorithmic unfairness and discrimination, bridging the notions adopted in social sciences, law, statistics, and artificial intelligence. (ii) To identify the ways in which algorithmic unfairness originates and spreads in different social contexts, affecting individuals and groups, and particularly to identify the cases in which algorithmic unfairness leads to prohibited discrimination. (iii) To analyse the ways in which the law currently addresses algorithmic discrimination and propose appropriate measures to implement or upgrade the existing regulatory framework. (iv) To examine the way in which technologies can promote fairness and support detecting and countering algorithmic unfairness and discrimination, in particular with regard to the assessment of asylum requests. By identifying and remedying algorithmic unfairness and discrimination, EquAl will contribute to preventing and mitigating harms to individuals and groups and favour the law-abiding deployment of AI. EquAl is premised on the fast-growing application of AI techniques for the purposes of prediction, evaluation, and decision making. Algorithmic approaches have the potential to transform many aspects of the economic and social life, delivering cost effective solutions, increasing the equity, efficiency, controllability and precision of decision-making processes. However, they may also lead to new and more subtle, opaque, and resilient forms of unfairness and discrimination. Some discriminatory effects have been already addressed by case-law in Europe and beyond, and some proposals exist to regulate aspects of automated decision-making, but no comprehensive regulatory framework exists yet. EquAl aims to place Italian legal research at the forefront in the domain of algorithmic fairness and non-discrimination, by: (a) delivering new insights on the specific nature, functioning, and evolution of fair and unfair instances of algorithmic decision-making; (b) evaluating existing anti-discrimination technologies and developing new methods to detect instances of unfairness in human and automated decisions and protect vulnerable individuals; (c) providing ethical and legal guidance and (d) supporting public bodies, NGOs and local communities, in particular, in the examination of asylum applications. EquAl’s contribution is crucial to enhance interdisciplinary cross-fertilisation, since currently different criteria and terminologies are used in debating algorithmic fairness and non-discrimination by different research communities (legal scholars, sociologists, computer scientists, statisticians), and to ensure that the corpus of EU and Italian anti-discrimination law, regulations, and case-law can be effectively applied in the algorithmic domain.

Legal Analytics for Italian Law (LAILA)

Sat, 01 Jan 2022 00:00:00 +0000

The project regards the application of Legal Analytics methods to a vast and heterogeneous set of legal information: legislations, contracts, and judgments. The purpose is the application of Artificial Intelligence, Machine Learning, and Natural Language Processing to extract legal knowledge, infer relationships, and produce data-driven forecasts.

Analytics for DEcision of LEgal cases (ADELE)

Fri, 01 Jan 2021 00:00:00 +0000

Project ADELE is premised on the ongoing paradigm shift towards cognitive computing and human-centered AI which is transforming many socio-economic activities, including justice. The project applies legal analytics (LA) – a blend of data science, machine learning, and natural language processing techniques – to judicial decisions. It aims to develop methods to extract knowledge and engage in outcome predictions and there build a pilot tool to support legal research and decision-making processes in the judiciary.

automated CLAUse DETectEr (CLAUDETTE)

Sun, 01 Jan 2017 00:00:00 +0000

CLAUDETTE is an interdisciplinary research project hosted at the Law Department of the European University Institute. The research objective is to test to what extent is it possible to automate reading and legal assessment of online consumer contracts and privacy policies, to evaluate their compliance with EU’s unfair contractual terms law and personal data protection law (GDPR), using machine learning and grammar-based approaches.