Developing AI for the legal sector: innovation and collaboration at the University of Liverpool

<span>Photograph: David Lees/Getty Images</span>
Photograph: David Lees/Getty Images

From smartphones with facial recognition to Amazon product recommendations, automated candidate screening and chatbots such as ChatGPT, artificial intelligence (AI) is transforming the way we live and work. And that includes driving real change in the legal profession too.

Pioneering, legally-grounded AI models have been developed by researchers at the University of Liverpool that can advise lawyers as to whether a case is likely to be successful or not. The suite of technologies – which are already being used in the real world – are proving to be transformative for the legal sector, spelling an end to long hours searching through documents and looking up case precedents.

A feature of Liverpool’s research expertise is developing a computational model of argument (CMA) that delivers explainable AI. Such techniques are used to capture knowledge about a body of case law or legislation and advise legal professionals on the applicability of the domain knowledge to a set of facts. Working with UK law firm Weightmans LLP, University of Liverpool researchers created a prototype where testing showed the explainable AI tools had very high levels of accuracy in matching outcomes in historic industrial deafness cases.

Researchers continue to collaborate with the law firm, which is testing the prototype AI solution in three different fields of law. And they have continued to innovate – creating bespoke AI solutions.

AI as a decision aid
Early evidence of successful deployment of new legal AI technologies can be seen at Fletchers, one of Britain’s biggest medical negligence firms, which has deployed a co-created model in its initial claims work and is now in the process of developing it further. Staff used to spend long hours weighing up the chances of success for clients pursuing claims for serious or life-changing injuries. But today they can tell in seconds whether a client has a strong case thanks to the creation of a revolutionary digital assistant or “explainable AI system” that acts as a decision aid.

Already, it has transformed working practices at Fletchers – increasing efficiency, reducing costs and turnaround times, with the result that more clients have access to justice – says Dan Taylor, the company’s director of integration.

“For us, AI is much more than a buzzword. We can say it has transformed our ability to serve clients and open up that access to justice for seriously injured people,” Taylor says.

The new AI model is the latest tool to be developed by researchers from the University of Liverpool’s school of electrical engineering, electronics and computer science, which has been producing innovative research in the field of AI and the law since the 1980s. Indeed, the strength of its research has established the dean of the school, Prof Katie Atkinson, as a thought leader in the field.

However, it is only relatively recently that the university transferred the new technologies to the real world, thanks in part to the Knowledge Transfer Partnership (KTP) scheme funded by Innovate UK – a mutually beneficial grant-funded partnership that connects businesses with an innovation idea to the academic expertise needed to deliver it, while being cost-effective.

Over the past decade, the university seized on interest in the lawtech market by working with businesses to deliver pioneering ways to extract and reason with knowledge from complex legal documents.

It is funding from the KTP scheme that meant the university could work with Fletchers to develop support tools for its specialist area of law using the university’s AI research. The grant, funded by Innovate UK, allowed two KTP associates to be embedded within the firm for three years under the supervision of academics.

For Fletchers, the partnership was a “perfect marriage”, says Taylor, explaining that the company spent a year working with a large multinational tech firm investigating if it could adapt an AI engine for them, capable of risk-assessing potential cases. Finding it could spend millions of pounds with no guarantee of results, Fletchers turned to the University of Liverpool.

“I’ve never stopped thinking about how lucky we were to have those guys on our doorstep,” says Taylor.

“We get thousands and thousands of inquiries each month,” he says. These cover childbirth injuries, misdiagnoses, oncology and plastic surgery accidents, lots of repeatable things that are recorded in tens of thousands of decisions in the past. “The algorithm can cross-reference past decisions in ways that a human being just cannot do.

“The digital assistant is used every day in businesses, freeing up lawyers to spend more time with more clients. Lawyers interviewing clients can now cross-reference their data against thousands of similar cases and tell them whether they have a decent case and their chances of success within minutes.”

Shot of a group of programmers working together on a computer code at night
Lawyers interviewing clients can cross-reference data against thousands of similar cases. Photograph: Jay Yuno/Getty Images

Although the KTP ended in 2018, Fletchers has taken on the two KTP associates as full-time employees to develop the AI model for different parts of the case system. Throughout, transparency has been a key part of its development, with data from the AI model audited by staff.

“Working with the University of Liverpool through the KTP made us smarter and helped us deliver faster access to justice for a wider range of seriously injured clients. The grant funds and the expertise turbocharged our innovation strategy beyond where we thought we could go, and made us think very differently about how you provide an excellent service at a critical time in their lives,” Taylor adds.

None of this could have happened without the KTP scheme, which has proved to be a vehicle for change for a number of companies without in-house legal AI expertise. Over the years, the university has been involved in developing different AI tools for other companies in the region, including Wirral-based Riverview Law (now part of accountancy firm EY), and more recently, conveyancing firm My Home Move Conveyancing.

Prototype tools to check admissibility
Research continues apace on developing tools that enable faster and more consistent decision-making. Atkinson, who was former president of the International Association for AI and Law, has led a team that has built a prototype AI tool for lawyers to check on the admissibility of cases submitted to the European Court of Human Rights. Meanwhile, the university is also working with Barclays Eagle Labs and a number of different law firms in supporting lawtech startups.

Urging business to come to the university with any problems they feel could be solved by AI technology, Atkinson says a range of AI solutions are applicable to the legal sector that are grounded in mature academic research.

“AI technology takes time and resources to develop, but the positive impact on efficiency, accuracy and reduced case processing time has been found to significantly outweigh the initial expenses in the projects we have undertaken.

“People don’t want to be waiting years to get their cases heard and get access to justice. The law firms we work with want to provide excellent customer service, and you do that when you get a quick and transparent assessment of your cases.”

If you’d like to find out more about how a KTP grant can help you tap into the AI expertise at the University of Liverpool, and take your business to the next level, head here