Interview
Royal Institution Christmas Lectures 2019: How can we all win?
Professor Katie Atkinson discussed the opportunities and challenges of using AI in the legal sector as part of the prestigious Ri Christmas Lectures.
What is it about?
Legal professionals process extensive amounts of lengthy documents, even for routine legal work. What if AI could process these documents instead and offer guidance on decisions in legal cases, based on previous precedents? This would streamline administration, drive efficiencies, and deliver swifter access to justice. At the same time, it would allow legal professionals to concentrate on high-level tasks such as action planning, strategy planning, and client support activities that require social intelligence and human contact.
The University of Liverpool has seized upon a recent and unprecedented interest in AI by legal practice and is delivering a step change in the lawtech market by pioneering ways to extract, and reason with, relevant knowledge and data from complex legal documents. The knowledge is built into a computational model of argument (CMA) to develop decision-support tools that enable more consistent and faster decision making.
CMA tools have been demonstrated as highly accurate in replicating the outcomes of closed court cases in a variety of well-studied domains, reaching 100% success in certain areas of law. Other AI tools under development provide decision support that can recommend choices while displaying the arguments and justification process, guiding public sector judiciary and commercial legal professionals towards informed actions.
For example, CLIEL (Commercial Law Information Extraction based on Layout) applies AI techniques such as natural language processing (NLP), data mining and machine learning to provide a fast, flexible and scalable system for extracting information from legal documents related to commercial law. The ANGELIC methodology is another development by Liverpool academics, which enables AI programs to be designed to capture knowledge about a body of case law and advise on acceptable arguments in a case.
Researchers at Liverpool are working with a variety of law firms to apply different AI techniques to different legal tasks in order to move legal services into the next generation, one augmented by AI.
Why is it important?
AI is transforming the way we work, and the legal sector is no exception. AI technology development takes time and resources, but the positive impact on efficiency, accuracy and reduced work hours can significantly outweigh the initial expenses. In addition, time savings enable earlier risk mitigation that can lead to better outcomes for businesses and their customers.
The University of Liverpool has been innovating in AI and law since the 1980s, interacting closely with the International Association for AI and Law (IAAIL). Professor Katie Atkinson and Professor Trevor Bench-Capon from the Department of Computer Science have both served as president of IAAIL, leading to fruitful collaborations with local and national law firms and the development of bespoke AI tools.
The strength of research conducted by the University of Liverpool has established Prof Atkinson as a thought leader in AI and law. Atkinson offers guidance to the Lord Chief Justice’s AI Advisory Group and who regularly engages legal professionals and the broader public with the ethical issues relating to AI and law.
Perspectives
There is currently a lot of interest in AI within the legal community and we are now seeing the deployment of AI in legal work. One strand of research that we undertake at Liverpool is specifically focused on ‘explainable AI’, which is a very important issue to be addressed to build trust and confidence in the use of AI by legal professionals and their clients. Our models of legal decision making have explainability features because we use computational models of argument. Humans use argumentation to debate and discuss what to believe and what to do, so we are building formal models that can be turned into computer programs for automated argumentation. These programs can take a set of arguments about a case and determine which are the acceptable ones and why, based on precedent cases. We also build in reasoning about the social values that are promoted by different legal decisions. We’re very keen to transfer this research into practice so have been working with a number of law firms to deploy AI technologies in their business operations. We aim to help the firms make faster, more consistent decisions with the help of transparent AI tools, and ultimately deliver justice for the law firms' clients.
