Recap panel Law en Tech lab

On September 13, the Law and Tech Lab hosted a panel event on “Bridging the gap between Logic and NLP-based Approaches for Automating Regulatory Compliance”.

The event was organised by Rohan Nanda, Aurelia Tamò-Larrieux, Hellen van der Kroef and Gijs van Dijck, and hosted renowned experts in the field: Monica Palmirani from the Department of Legal Studies at the University of Bologna; Réka Markovich from the Department of Computer Science at the University of Luxembourg; Ilias Chalkidis from the Department of Computer Science at the University of Copenhagen; and Luigi Di Caro from the Department of Computer Science at the University of Turin.

The event addressed essential issues within regulatory compliance. Regulatory compliance refers to an organization’s conformity to relevant laws and regulations. The role of regulatory compliance software is to enable economic operators to understand what regulations, policies, and obligations are applicable to them. The majority of existing software solutions still stick to the traditional approach of providing expert content in the form of advice and commentaries. However, the tremendous pressure of the continuously increasing unprecedented number of domestic, European Union (EU), and international regulations and case law is challenging this model. Regulations and related data are often in the form of open data accessible on the web. However, this availability does not solve the problem: regulations are split in different silos, even internally within countries. The identification, extraction, and formalization of legal norms for automated compliance checking remain a challenge for the Computational Law field.

To address these challenges, Artificial Intelligence (AI)-based solutions have been developed to automate specific steps of legal compliance. These AI solutions fall into two broad categories: Logic and natural language processing (NLP)-based technologies for computational analysis of law resulting in machine-readable or executable legal norms. The NLP-based approaches are used to extract information from legal documents and formalize it in a machine-readable format. The logic-based approaches then use this formalized knowledge with reasoning tools for enforcing legal norms and assessing compliance. However, the success of the recent language models (PaLM and GPT-3) in reasoning and natural language inference and understanding tasks has increased the possibility of standalone NLP-based systems for automated compliance checking.

The objective of our event were to bridge the gap between the logic and NLP-based approaches for automating legal compliance and discuss: The strengths and limitations of both logic and NLP-based approaches for legal compliance, how these approaches can complement each other, and the future directions of research for automating legal compliance with AI.