PRIN2017 - LAILA: Legal Analytics for Italian Law

The project (LAILA: Legal Analytics for Italian LAw) addresses the application of methods of legal analytics (LA) to a vast and diverse set of legal information: legislation, case law, and empirical legal data. It studies the use of analytics—a mix of data science, artificial intelligence (AI), machine learning (ML), natural language processing, and statistics—in the legal domain, to extract legal knowledge, infer undiscovered relationships, and engage in data-driven predictions.  LAILA is premised on the data-and-AI paradigm shift that has taken place in the last two decades: AI technologies, as applied to vast datasets (big data), have already transformed many aspects of our economic and social life, delivering many useful applications, but also causing some serious drawbacks (surveillance, monopoly power, manipulation of individual and social choices, etc.). The study and the practice of the law are also being affected as new powerful methods become available for legal cognition and practice.

In this context, LAILA will aim to place Italian legal research and practice at the forefront in the use of legal analytics, by testing and developing technologies, providing ethical and legal guidance, and delivering new insights on the functioning and evolution of law. LAILA’s contribution is crucial to this objective: while vast repertoires of digital resources on Italian law are already available, LA tools and methods for adequately processing such resources are still missing. In fact, most studies and products for legal analytics concern common law systems (mainly the US) and the English language.

LAILA has two main targets:

1. To apply, assess, refine, and develop technologies for Legal Analytics (LA). This includes:

• Applying analytics technologies—including supervised, semi-supervised, and unsupervised learning—for the following purposes: ontology building, classification of legal documents, analysis of legislation, analysis of case law, extractions of “massime” (rationes) and principles, question-answering, and prediction of trends and decisions.
• Refining and developing LA technologies, taking into account the peculiarities of the Italian law and language. In particular, we shall combine LA on English and Italian texts, using existing multilingual repositories; apply LA to annotated legal document; and combine LA methods with the logical representation of legal knowledge.

2. To provide methodological analyses and guidelines for LA. This includes:

• Reviewing the capabilities and limitations of current LA technologies, and identifying risks and mitigations;
• Providing ethical and legal guidelines, specifying how to deploy LA for different purposes (classification, analysis, prediction, etc.), while complying with law and ethics and preventing any risks of adverse effects (biases, unfairness, etc.).